Classificatory Particles in Kilivila
OXFORD STUDIES
IN ANTHROPOLOGICAL LINGUISTICS William Bright, General Editor
C...
13 downloads
970 Views
16MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Classificatory Particles in Kilivila
OXFORD STUDIES
IN ANTHROPOLOGICAL LINGUISTICS William Bright, General Editor
Classificatory Particles in Kilivila Gunter Senft
Classificatory Particles in Kilivila
Gunter Senft
New York Oxford OXFORD UNIVERSITY PRESS 1996
OXFORD UNIVERSITY PRESS
Oxford New York Athens Aukland Bangkok Bombay Calcutta Cape Town Dar es Salaam Delhi Florence Hong Kong Istanbul Karachi Kuala Lumpur Madras Madrid Melbourne Mexico City Nairobi Paris Singapore Taipei Tokyo Toronto and associated companies in Berlin Ibadan
Copyright © 1996 by Gunter Senft Published by Oxford University Press, Inc. 198 Madison Avenue, New York, New York 10016 Oxford is a registered trademark of Oxford University Press All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data
Senft, Gunter 1952Classificatory Particles in Kilivila / Gunter Senft p. cm, - (Oxford Studies in Anthropological Linguistics) Includes bibliographical reference and index. ISBN 0-19-509211-2 1. Kiriwinian language-Particles 2. Kiriwinian language-Classification 3. Trobriand Islands (Papua New Guinea-LanguagesGrammar) I. Title 11. Series PL6252. K5S38 1995 499'. 12-dc20 Printing (last digit) : 9 8 7 6 5 4 3 2 1 Printed in the United States of America on acid free paper
pela minana sinebada namanabweta nakabitam ula kwava nabweliguee
BARBARA
This page intentionally left blank
. . . Cassirer 1923 behandelt im ersten Teil seiner Philosophie der symbolischen Formen die menschliche Sprache, wobei auch die "Grundrichtungen der sprachlichen Klassenbildung" besprochen werden... Das Wort "Philosophie" mochte vielleicht Namen, wie die eines MAUTHNER und HOOGVLIET und andere mehr in Erinnerung rufen, hier aber ganz zu unrecht. Dieses Buch ist reich an Gedanken und Einsichten, auBerst suggestiv und lehrreich. Es sei nicht die Aufgabe der Sprachphilosophie "die verschiedenen Formen der Begriffs- und Klassenbildung, die in den Einzelsprachen wirksam sind, zu beschreiben, und sie in ihren letzten geistigen Motiven zu verstehen... Die Wege, die die Sprache hier einschlagt, sind so vielfaltig verschlungen und so dunkel, daB es nur durch die genaueste Versenkung und durch die feinste Einfuhlung in das Detail der Einzelsprachen gelingen kann, sie allmahlich zu erhellen. Denn gerade die Art der Klassenbildung macht ein wesentliches Moment jener "inneren Form" aus, durch welche sich die Sprachen spezifisch voneinander unterscheiden". Man sieht, daB sich CASSIRER visionarer Fliige in das Reich der erphantasierbaren Moglichkeiten enthalt und dauernde Fuhlungnahme fordert mit der reichen Mannigfaltigkeit der sprachlichen Tatsachen. Aber wenn auch die verschiedenen Klassifikationssysteme in verschiedenen Varianten wechselten, so konnten doch in dieser groBen Mannigfaltigkeit "gewisse allgemeine Gesichtspunkte" entdeckt werden. Zweifellos sei es moglich, diese Gesichtspunkte so anzuordnen, "daB man dabei jenen standigen Fortgang vom "Konkreten" zum "Abstrakten", der die Richtung der Sprachentwicklung uberhaupt bestimmt, als leitendes Prinzip benutzt". Aber, so fugt CASSIRER . . . hier richtig hinzu, man durfe nicht vergessen, "daB es sich hier nicht um eine zeitliche, sondern um eine methodische Schichtung handelt, und daB demnach in einer gegebenen historischen Gestalt der Sprache die Schichten, die wir hier gedanklich zu sondern versuchen, neben- und miteinander bestehen und sich in der mannigfachsten Weise iibereinander lagern konnen". Gerlach Royen (1929:254-255) . . . wer auf der Studierstube ein System zimmert, ohne es der Welt anzupassen, der lebt entweder seinem System all Augenblick schnurstracks zuwider, oder er lebt gar nicht. Jakob Michael Reinhold Lenz (1774)
This page intentionally left blank
Preface We shall deal in this article with a single phenomenon, namely, the classificatory formatives in the language of Kiriwina, Trobriand Islands, an archipelago lying due north of the eastern end of New Guinea. Bronislaw Malinowski (1922:37)
Classification is certainly a basic fact of life. That classification abilities are necessary for the survival of every organism is an important insight of biology. Human beings classify consciously and unconsciously—and subconsciously, of course—in all situations. When we confront a scientific problem, we try to solve it by first classifying the various parts of the problem. Thus it is not surprising that the history not only of philosophy but of all branches of science is also the history of how these sciences have classified their research subject. Classification always implies selection because, as Koestler (1978:201) stated,
(our) minds would cease to function if we had to attend to each of the millions of stimuli which—in William James's classic phraseconstantly bombard our receptor organs in a "blooming, buzzing confusion." Thus the nervous system and the brain itself function as a multilevelled hierarchy of filtering and classifying devices, which eliminate a large proportion of the input as irrelevant "noise," and assemble the relevant information into.coherent patterns before it is represented to consciousness.
If we want to communicate about this perceived, classified, and filtered input, we have to classify once more: We have to transform this input into classes and categories provided by the systems that organize our communicative verbal and nonverbal faculties. With our systems of language and gesture, we again classify and filter on various levels while communicating. On one of these levels, we decide, for example, on the grammatical structure we want to use to refer to what we want to communicate about; on another level, we also have to classify the referents of our communication. Linguistics is the science that tries to describe, illuminate, and explain the processes of classification that are relevant for communication—and this goal is the reason that many
X
Preface
linguists find their discipline so fascinating. Indeed, human beings have developed a number of different linguistic techniques to apprehend the world, providing an enormous database for the analysis of this problem. This book deals with a classification technique used by the speakers of an Austronesian language, Kilivila. In August 1982, I entered for the first time the district that was to be the future site of my linguistic fieldwork—the Trobriand Islands in Milne Bay Province of Papua New Guinea. I did so with, I suppose, almost the same feelings of intense interest and suspense that Malinowski felt (Malinowski 1922:51). However, I had the enormous advantage of having read Malinowski's masterpieces on the Trobriand Islanders' culture. Nevertheless, there is a difference between having read about a completely foreign world and actually confronting it. The aim of my research project was to describe and explain aspects of ritual communication. To achieve this aim, it was necessary to acquire a certain competence in the language, of course. Since there was no existing grammar of the Trobrianders' language, writing a grammar was one of the prerequisites (G. Senft 1986:3-5). Among the few linguistically reliable sources on Kilivila was Malinowski's fascinating article "Classificatory Particles in the Language of Kiriwina," published in 1920. Working on this system of classifiers became one of my preoccupations as soon as I was able to master the language. After fifteen months of field research, I returned to Germany with a great deal of data on classifiers, knowing that I would concern myself for a long time with the analysis of these data and with an attempt to understand this system and its importance for the language as well as for the culture. This concern is documented in a number of publications (G. Senft 1983, 1985a, 1985b:133-134, 1986:68-72, 1987a, 1989, 1991c, 1993). In this book, I present the results of almost ten years' work on the system of classificatory particles in Kilivila. However, I am not at all sure whether this means that I am really "through" with my Trobriand friends' classifier system. The work is offered as a contribution to the research on classifiers and classifier languages. It is an empirical work, based on data gathered during a period of fifteen months of field research in 1982 and 1983 and four months of field research in 1989. It is extremely data-oriented and emphasizes the use of classifiers in social contexts. The aim of this work is to describe (1) the functions of the system of Kilivila classifiers; (2) the acquisition of the classifier system; (3) the inventory of Kilivila classificatory particles (produced in actual speech); (4) the processes of language change that affect the system; and (5) the semantics of the Kilivila classifier system. The introductory chapter—based on a broad survey of the relevant literature—gives a general definition of the concepts "classifier languages" and "classifiers" and presents the aims and methods pursued in traditional studies on the classifier systems of classifier languages in general and on the Kilivila language in particular. Chapter 2 presents a description of the
Preface
xi
grammatical and discourse functions of the system of classificatory particles (CPs) in the Kilivila language. The third chapter is the central chapter of this book. After a description of the aims and methods used in data collection, the CP data gathered in a specially developed elicitation test, as well as CP data documented in my overall corpus of Kilivila speech, are presented in detail. First, the CP production data are presented and preliminarily interpreted separately for each of the five age groups into which I divided my consultants. This data presentation is meant to give the reader a first impression and a general account of CP production in each of the five age groups. Second, the data gathered in the CP elicitation test are presented in two ways: (1) the CP types and the respective tokens produced by all consultants during the elicitation test are given, including a description of which consultant produced how many tokens of the respective CP type; and (2) the CP types produced by all consultants in the elicitation test are given, allocating the tokens of the CP type produced to the CP type or types actually expected. Finally, the CP types and tokens documented in my overall corpus of Kilivila speech data are presented. This kind of data presentation serves as an empirical, checkable basis for the analyses necessary to reach the aims of this study. I am aware that this large-scale presentation requires space; however, it allows the critical reader to check (and countercheck) all the analyses and all the inferences made using these data. Moreover, it is hoped that this large-scale data presentation makes it easier for the reader to understand the train of thought that starts with the empirical datum elicited (on the basis of theoretical reflection and ideas presented in the relevant literature) and leads to the description, interpretation, and evaluation of this empirical datum, resulting in theoretical conclusions. Following the data presentation, the results of the CP research are given. First, the question of how the CP system is acquired by Trobriand children is considered. Here, all but one of the many side issues of this complex question are answered. The subquestion of why the individual CPs are acquired in the order found is the last question addressed in this chapter, because the answer requires the information provided by the answers to the other two main questions raised in the study. The question of how the individual CP types are produced in actual speech is considered next. The third and last question answered in chapter 3 is that of the semantic domains constituted by the CP system. The headings of the sections and subsections in chapter 3 should make it easy for readers to find the answers to aspects of this study that may be of special interest to them. Chapter 4 presents the results of a restudy I did in the Trobriand Islands in 1989, in which I reconsidered the results of my analyses of the Kilivila CP system based on data gathered in 1982 and 1983. Chapter 5 is an excursus; it discusses—with all necessary caution—possible interdependencies between language, culture, and cognition based on an analysis of the CP system.
xii
Preface
Chapter 6 summarizes the results of the study and presents a "network" model for the description of classifier systems. Appendix A lists the individual consultants used in gathering CP data in 1982 and 1983; appendix B lists which consultant produced which CP in what text category in my overall corpus of Kilivila speech data; appendix C lists the consultants I worked with during my restudy in 1989; and appendix D presents some hypotheses on the origin of classifiers. I would like to mention here two concepts that are central in understanding the research and the arguments that I present. First, I speak of the "system" of CPs in Kilivila and of "systems of classification" in general. I do this not only because I follow the general usage in the literature, but especially because I am convinced that my analyses of the CPs in Kilivila justify the use of the term "system." One of the aims of this research is to describe a set of elements—that is the set of CPs in Kilivila—and the set of relations existing among these elements (Klaus 1968:634). I try to describe the internal order among linguistic elements and to present the functional relation that can be found on various levels of description and in relation to social and other subsystems (Bussmann 1983: 489). Second, I speak of "referents" and of the "act of referring." A referent is an object or a fact in the extralinguistic reality to which noun phrases as verbal signs "refer." By "act of referring," I understand on the one hand the verbal reference to language-internal and language-external contexts and on the other hand the relation between the verbal expression (name, word, etc.) and the object in the extralinguistic reality to which the expression refers (Bussmann 1983:428). I would like to end this preface with a warning to my readers: much of this book contains rather dry linguistic descriptions. However, I hope that the results presented help readers forget some of these long hauls and that they may get at least an impression of the fascination I experience in dealing with the complex linguistic phenomenon of classifier systems in general and in particular the system of classificatory particles in Kilivila. June 1995 Nijmegen, The Netherlands
G.S.
Acknowledgments
During my work in Tauwema in the Trobriand Islands and during the writing of this book, I received the help of many people and institutions. First, I want to thank the Deutsche Forschungsgemeinschaft (DFG: Ei-24/ 10-1-5; Se-473/2-1-2), especially Ursula Far Hollender and Manfred Briegel; the Research Unit for Human Ethology of the Max-Planck-Society; and the Cognitive Anthropology Research Group at the Max-Planck-Institute for Psycholinguistics for their support during my field research. I am also indebted to Penelope Brown, Bernard Comrie, Volker Heeschen, Stephen C. Levinson, Bernd Nothofer, Roland Posner, and Barbara Unterbeck for their critical reading of first drafts of this monograph; their extensive comments were extremely valuable to me in the preparation of this book. I would like to thank William Bright and two anonymous referees for their insightful comments, and I thank the team at Oxford University Press for their helpfulness and expertise. I want to thank the National and Provincial Governments in Papua New Guinea and Milne Bay Province for their assistance with, and permission for, my research projects. I also thank the Departments of Linguistics and Anthropology of the University of Papua New Guinea, the Institute of Papua New Guinea Studies, the Council of Chiefs of the Trobriand Islands, and the Embassy of the Federal Republic of Germany in P.N.G. for their support of my research. I especially thank the people of the Trobriand Islands, and above all the inhabitants of Tauwema and my consultants for their hospitality, friendship, and patient cooperation. Without their help, none of my work on the Kilivila language and the Trobriand culture would have been possible. Although I typed the manuscript myself, finishing touches have been made by Edith Sjoerdsma and Gertie de Groen. All the problems I had with the various computers I used were competently solved by Gertie de Groen, Caroline Rek, Herbert Baumann, and Karl Grammer. Many thanks to all of them. A German version of this manuscript was accepted as my Habilitationsschrift at the Fachbereich 1 of the Technische Universitat Berlin. I would like to thank all the people involved in the process of my Habilitation for their support and for helpful comments on my research. Last, but by no means least, I want to thank my wife for her patience and perseverance in discussing my research with me, criticizing it, and living and working with me in the field. I dedicate this monograph to Barbara Senft.
This page intentionally left blank
Contents
Chapter 1 Introduction 1.1 1.2 1.3 1.4 1.4.1 1.4.2
What Are Classifier Languages? What Are Classifiers? Structure and Function of Classifier Systems Methods and Aims of Traditional Classifier Studies—A Brief Survey Studies on Classifier Languages Studies on Kilivila
3 3 5 9 11 11 14
Chapter
2 Classificatory Particles in Kilivila: Grammatical and Discourse Functions 2.1 2.2 2.2.1 2.2.2 2.2.3 2.3
Morphological Relevance Functions of CPs in Kilivila Referential Function—Concord Nominalization, Plural Marking, Numeralization, and Verblike Expressive Functions Redundancy, Ellipsis, and Discourse Coherence Summary
16 16 18 18 20 20 22
Chapter 3 The System of Classificatory Particles in Kilivila 3.1 3.2 3.2.1 3.2.1.1 3.2.1.2 3.2.1.3 3.2.1.4 3.2.1.5
Aims and Methods The Data Interpretation of the Data by Age Group Age Group I Age Group II Age Group III Age Group IV Age Group V
24 24 31 32 32 41 52 62 73
xvi
Contents
3.2.2 Complete Lists of Data Used in the Analyses 3.2.2.1 CP Types/Tokens Produced by All Consultants During the Elicitation Test 3.2.2.2 CP Types/Tokens Produced Versus CP Types Expected During the Elicitation Test 3.2.2.3 CP Types and Tokens Documented in the Corpus of Transcribed Kilivila Speech 3.3 Results 3.3.1 Which Formatives Constitute the Kilivila CP System and What Is the Actual Occurrence of CPs in Recorded Tests? 3.3.2 How Is the CP System Acquired by Trobriand Children? 3.3.2.1 What Is the General Temporal Progress of the CP Acquisition Process? 3.3.2.2 Are There Any Differences in the CP Acquisition Process With Respect to the Production of Demonstrative Prounouns, Adjectives, and Numerals That Use CPs as Morphemes in Their Word Formation? 3.3.2.3 Are There Any Gender-Specific Differences in the CP Acquisition Process? 3.3.2.4 What Is the Order in Which the Individual CPs Are Acquired? 3.3.2.5 Are There Any Parallels Between Kilivila CP Acquisition and CP Acquisition Data for Other Classifier Languages? 3.3.2.6 Excursus: Are There Parallels Between the Kilivila Data and Acquisition Data With Respect to Demonstrative Pronouns, Adjectives, and Numerals for such Indo-European Languages as English and German? 3.3.3 What Is the Realization of the Individual CP Types in Actual Speech? 3.3.3.1 Can a Classification of the CP System Be Devised? 3.3.3.2 How Do the Consultants Realize the Individual CP Types and Can We Observe Any Processes of Language Change in Progress? 3.3.3.3 Why Are the CP Types Affected by Processes of Language Change, and Who Opposes and Who Fosters Language Change? 3.3.3.4 What Are the Possible Consequences of Language Change for the Grammar of Kilivila? 3.3.4 What Semantic Domains Are Constituted by the Kilivila CP System? 3.3.4.1 What Semantic Domains Are Constituted by the Described Subset of the Kilivila CP System? 3.3.4.2 What Are the Dynamics of the Semantic Domains Constituted by the CPs and What Rules Do Speakers Use in Their Production of a Certain CP? 3.3.4.3 A Flashback: Why Are the Individual CPs Acquired in a Certain Order?
86 87 134 161 171 171 180 180 181
182 183 193
195
197 198 202 227 236 236 237 242 290
Contents
xvii
Chapter 4 On the Validity of Some of the Presented Results: Six Years Later—A Restudy
295
Chapter 5 Excursus: Language, Culture, and Cognition?
312
Chapter 6 Closing Remarks: Using Network Models to Describe Classifier Systems Appendix A Appendix B
Appendix C Appendix D References Index
Consultants: 1982/1983 Study Number of CP Tokens Produced for Each CP Type by Text and Word Class for Consultants in Corpus of Kilivila Speech Data Consultants: 1989 Restudy Some Speculations on the Origin of Classifiers
323 330 333
339 352 355 371
This page intentionally left blank
Classificatory Particles in Kilivila
This page intentionally left blank
1
Chapter
Introduction Kilivila (also known as Kiriwina and Boyowa) is one of the 40 Austronesian languages spoken in the area of Milne Bay Province in Papua New Guinea. It belongs to the "Papuan Tip Cluster" (Capell 1976:6, 9; Ross 1988:25-27), and typologically is classified as having verb-object-subject (VOS) word order (G. Senft 1986:107-112). The Kilivila language family comprises Budibud (also known as Nada), Muyuw (also known as Murua), and Kilivila. Kilivila has about 25,000 speakers, most of whom live on the Trobriand Islands. Bronislaw Malinowski's ethnographic work on the Trobriand Islands has made this area rather well known, even outside the field of anthropology. In fact, Malinowski first reported the phenomenon that is the subject of this book. As a result of his classic article, "Classificatory Particles in the Language of Kiriwina" (Malinowski 1920), Kilivila is known to linguists as a so-called classifier language (Allan 1977:286-288). The present chapter first defines the concepts "classifier" and "classifier language". It then gives an overview of the structure and function of classifier systems and the methods and aims of traditional classifier studies, including those on classifier languages in general and those on Kilivila in particular.
1.1 WHAT ARE CLASSIFIER LANGUAGES?
John Lyons (1977b:463) notes an important feature of classifier languages: Languages which grammaticalize the distinction between entitydenoting nouns and mass-denoting nouns tend to draw a sharp syntactic distinction between phrases like "three men" on the one hand, and "three glasses of whisky," on the other. Classifier languages do not: they treat enumerable entities and enumerable quanta in much the same way. 3
4
Classificatory Particles in Kilivila
Classifier languages also show the following three characteristics (Allan 1977:286-8): 1. They have a system of classifiers. 2. They follow the almost universal principle that "A classifier concatenates with a quantifier, locative, demonstrative or predicate to form a nexus that cannot be interrupted by the noun which it classifies" (Allan 1977:288; but see also Adams 1989:12, 24). 3. They belong to one of the following four classifier language types: numeral, concordial, predicative, or intralocative. Classifier languages are distributed around the world and are found in such different language families as Malayo-Polynesian, Mon-Khmer, Austro-Asiatic, Sino-Tibetan, Altaic, Dravidian, and Indo-Aryan (see, among others, Adams and Conklin 1973:9; Foley 1968:77-91; Greenberg 1975:18).2 Numeral classifier languages are considered the paradigmatic type, and Kilivila belongs to this class.3 Languages with numeral classifiers differ from other languages primarily with respect to the following characteristic feature: in counting inanimate as well as animate referents, the numerals obligatorily concatenate with a certain morpheme, which is the so-called classifier.4 This morpheme classifies and quantifies the respective nominal referent according to semantic criteria (see Serzisko 1980:1, 1982a:147; Hundius and Kolver 1983:166).5 Moreover, 1 For further examples where one noun takes over similar classifying functions with respect to a second noun in the noun phrase, see Adams and Conklin (1973); Allan (1977:292-293, 301); Burling (1965:260); Clark (1976:449-450); Denny (1976:129); Greenberg (1975:22); Hla Pe (1965:167), Hoa (1957:125-126); Katz (1982); Kolver (1982c:162); Lee (1988:225, 242); Lobel (1986); Harweg (1987a); Jackendoff (1968); Parsons (1970); Plank (1984); Verhaar (1986, 1987:503, 1988:21-23); Zubin and Kopcke (1986); and Jacob (1965); for a formal logical discussion of count terms, mass terms, and quantification, see Lonning (1987); for the difference between noun classes and classifiers, see Dixon (1982: Part D:157-159); Silverstein (1986); Chin (1989, 27-29); but see also Gomez-Imbert's (1982) thesis on nominal classification in Tatuyo, a language that has features characteristic of classifier systems as well as features characteristic of noun class systems; for an interesting discussion of levels of semantic structuring in Bantu noun classification, see Spitulnik (1988). 2 Moreover, we find classifiers in American Sign Language (ASL); Egyptian hieroglyphics and Mesopotamian cuneiform use graphemic classifiers (see, e.g., Klima and Bellugi 1979: 13-15, 191-192; Kantor 1980; Newport and Supalla 1980; Kegl and Schley 1986; Supalla 1986; Rude 1986). 3 This book deals with the classificatory system of Kilivila; thus, I will not discuss the other three types of classifier languages mentioned. For a discussion of these types, see Allan (1977:286-287) and Craig (1986:3-4). 4 As the example of Kilivila shows, this morpheme may also be a so-called zero morpheme (G. Senft 1986:75; see also Mufwene 1980c:246). Moreover, due to processes of language change, it may well be that in some languages and some contexts the classifier is no longer obligatory, but becomes optional (see Chin 1989:4-6), Stolz (1991) presents data for classical Nahuatl, where the classifier is optional (see also Adams 1989). 5 What Seiler and his research team call Das sprachliche Erfassen von Gegenstanden (and living beings, I would add) by means of numeral classification is, by the way, just one of nine different techniques this linguist defines as Sprachhandlungsprogramme describing the dimension "apprehension" within the framework of the so-called Kolner Universalienprojekt (see Seiler 1982).
Introduction
5
in numeral classifier languages, we find classifier morphemes in anaphoric (see, e.g., Downing 1986) and deictic expressions. Therefore, the term "numeral classification" is somewhat inaccurate (see also Asmah 1972:90; Berlin and Romney 1964:79; Unterbeck 1990b:90). Nevertheless, I will adhere to this technical term as it is used in the general linguistic literature (see, e.g., Allan 1977:286; Becker 1975:114-115; Greenberg 1975:19; Haas 1942). Greenberg (1978:78) has answered the question of why such classifying systems apply only to nouns: . . . it is the noun par excellence which gives rise to classificational systems of syntactic relevance. It is not so much that the noun designates persisting entities as against actions or temporary states of persisting entities. It is that nouns are continuing discourse subjects and are therefore in constant need of referential devices of identification. As soon as we wish to talk about an action as such, we nominalize it; classification is a help in narrowing the range of possible identification.6 In sentence analysis, the syntactic description of the phenomenon of classification in numeral classifier languages is restricted to the noun phrase level (see Kolver 1983:55).7 The order in which the three constituents— classifier, numeral, and noun—appear varies across languages; however, according to the principle formulated by Allan (1977:288), in general the noun does not interrupt the nexus formed by the numeral and the classifier (see also Adams and Conklin 1973:1; Adams 1989:12, 24). The morphological/syntactic role of classifiers is generally neglected in the literature in favor of their semantic functions. However, Goral's statement that "in [classifier] systems, syntax and semantics are inextricably intertwined" (Goral 1978:5; see also Friedrich 1970:381) holds for all classifier languages and confirms the definitions proposed for the concept classifier. 1.2 WHAT ARE CLASSIFIERS?
We have defined classifiers as morphemes that classify and quantify nouns according to semantic criteria.8 Because of the twofold function of classifiers, Serzisko (1980:7)—following Hla Pe (1965:166) and Bloomfield (1933:237)— proposes the generic term "Numerativ" to denote the "obligatorische Konstituente in Quantifizierungskontexten" (see also Hundius and Kolver 6
See also Mufwene 1980a:1025 and Broadfield 1946:25. On the other hand, Schafer (1948:413) states the following about classical Chinese: "Classifiers are also used with numbers accompanying verbs 8 For general definitions of classifiers, noun classifiers, and nominal classifiers, see Allan 1977:285; Becker 1975:114-115; Benton 1968:137; Berlin 1968:20; Burling 1965:249; Denny 1979:97; Hoa 1957:124; see also the contributions in Craig (1986). Mufwene (1980a:1025) characterizes classifiers as "delimitative markers"; see also Mufwene 1984:200, 202. 7
6
Classificatory Particles in Kilivila
1983:167-169). The term numerative subsumes classifiers proper and quantifiers. Classifiers classify a noun inherently, that is, they designate and specify semantic features inherent to the nominal denotatum and divide the set of nouns of a language into disjunct classes (see Cholodovic 1954:49; Unterbeck 1990b:43). Quantifiers classify a noun temporarily, that is, they can be combined with different nouns rather freely and designate a specific feature of a noun that is not inherent to that noun. Thus, quantifiers are predicative (see Serzisko 1980:17, 68-69, 1982a:152; Berlin 1968:175; Friedrich 1970:397; Denny 1986:302-307). Contrary to one of Greenberg's (1975:25) language-universal postulates, not all but only the majority of nouns in numeral classifier languages lack a marking with respect to the category 'number'; in these languages, number is usually marked with nouns denoting persons (see, e.g., Barz and Diller 1985:170; Goral 1978:15; Miram 1983:36-37; Rausch 1912; Serzisko 1980:13, 48, 70; G. Senft 1986:45-46; Stolz 1991:18). Numeratives can take over the syntactic function of marking number for the nouns to which they refer. Referentially, nouns in classifier languages can be characterized as having generic reference (see Royen 1929:775). With their referential function, numeratives individualize nominal concepts; they can mark a noun as obligatorily nongeneric in reference (see H. Seiler 1982:6, 8; Serzisko 1980:15, 86-87; also Carpenter 1992:147). The functions that numeratives or classifiers fulfill are succinctly summarized by Adams et al. (1975:2): Besides their function in numeral noun phrases, classifiers in various languages function as nominal substitutes, nominalizers of words in other form classes, markers of definiteness, relativizers, markers of possession, and as vocatives; serve to disambiguate sentences; establish coherence in discourse and regularly mark registers and styles within a language.
These functions will be discussed in detail in chapter 2. So far we have differentiated classifiers and quantifiers (see also Adams 1989:3-5, 194) or, to use Lyon's (1977b:463) terms, "sortal classifiers" and "mensural classifiers" (see also Unterbeck 1990b:40). However, along with the definition of classifiers proper and quantifiers one generally finds a third category, the so-called repeaters. A repeater is a noun that serves as its own classifier, that is, the noun is repeated in the classifier slot. Hla Pe (1965:166) defines the terms classifier (proper), repeater, and quantifier as follows: A classifier is a word for an attribute of a specific object, some of which may have more than one; a repeater is the specific object itself or part of it, used as numerative; whilst a quantifier concerns
Introduction
7
itself with the estimating of things by some sort of measure—size, extension, weight or number especially of ten or multiples of ten. Repeaters are defined by Burling (1965:249) as "echo classifiers", Fischer (1972:69) calls them "identical classifiers", and Kolver (1982c:178, 183, 1979:34) characterizes them as "semantischer dummy"; finally, Goral (1978:33) defines repeaters as "autoclassifiers . . . filling a syntactic slot . . ." (see also Adams and Conklin 1974:3-4, 7; Benton 1968:116; Smith 1979:88; Carpenter 1992:132). In connection with this phenomenon, Lehmann (1979:169) hints at the possibility of studying this problem from a different point of view; he notes: ". . . a classifier can also function as an independent noun . . .". Whether this repeater category of numeratives, which probably has to be assigned to the category classifier, carries the principle of classification to the point of absurdity, as Kolver (1982c:178) claims, must be doubted after a closer inspection of the numeratives within the individual languages. Allan (1977:295) discusses the problem of repeaters with all necessary caution and offers some hypothetical answers to the question of why this category develops in languages.9 I shall not discuss this problem in detail here,10 but I want to emphasize that as far as I know there is no language described so far where all nouns can be used as repeaters to classify themselves. Thus, in all languages that have them, repeaters represent a relatively closed category; they form—in principle—a finite set of formatives. Moreover, most if not all classifier languages have, in the words of Lyons (1977b:461), at least one "semantically neutral classifier, which may be employed . . . with reference to all sorts of entities In many languages the semantically neutral classifier is restricted to non-personal, or even inanimate, entities . . . ." (see also Asmah 1972:95). This general classifier (see Serzisko 1980:24) has to be assigned to the category of classifiers, too (see also Schafer 1948:410-411). Finally, I want to note here that Malinowski (1920) does not differentiate between classifiers proper, quantifiers, and repeaters, but refers to these formatives as "classificatory particles". I will use Malinowski's general term (abbreviated as CP) for these formatives to pay tribute to the master of Trobriand ethnography.11 9 Jones (1970:2) states "It is interesting to speculate on the possibility that such usage arises from an inadequate supply of classifiers once their use becomes firmly established". See also Adams (1991) and Adams (1989:1). 10 Goral (1978:8) and Denny (1976:130) emphasize the strong affinity between nouns and numeratives (see also Schafer 1948:410-411). However, keeping in mind Dixon's (1968:114) admonition that we know very little about the development of such systems, I do not want to indulge in speculation on this topic here; but see appendix D of this volume and Dixon (1986); Mithun (1986); Delancey (1986); Demuth et al. (1986); Royen (1929:58, 63, 78, 141, 254, 266-268, 543, 705, 780); G. Senft (1993). 11 Royen (1929:iii, 37, 68, 185, 192, 305, 364, 889) emphasizes repeatedly that an interdisciplinary approach is not only necessary but also inevitable for an analysis of nominal classifier systems. See also Berlin et al. (1973:214). I want to mention for the sake of completeness that Berlin
(continued on page 9)
8
Classificatory Particles in Kilivila
Figure 1 summarizes this system—or classification—of CPs or numeratives and the technical terms found in the literature to refer to them.
General Terms for all Classifiers
Terms for the Category Classifier
Numeratives Classifjcatory particles (CPs) Classifiers Nominal classifiers Noun classifiers Zahleinheitsworter
Klassifikator Classifier proper True classifier Inherent state classifier Sortal classifier Repeater Reduplicative classifier Imitative classifier Echo classifier Identical classifier Semantically neutral/ Cannibalizing classifier Genereller Klassifikator Isolierter Klassifikator Inherent classification Inharente Klassifikation
Terms for the Category Quantifier
Quantifikator Quantifier Temporal state classifier Mensural classifier Action classifier Semelfactive classifier Metric classifier Temporary classification Temporare Klassifikation
Figure 1. Classification of Classifiers.
Introduction
9
1.3 STRUCTURE AND FUNCTION OF CLASSIFIER SYSTEMS
In classifier languages, nouns are classified and categorized according to characteristics of their referents. This classification is based on semantic principles and results in the ordering of objects, living beings, concepts, actions, and events. In other words, it leads to a categorization of all the nominal denotata, of all nominal "conceptual labels" (Hundius and Kolver 1983:182; see also Denny 1986) coded in the language. We can refer to the units of this classification as "semantic systems" (Denny 1979:97) or as "semantic domains" (Berlin 1968:34; Tyler 1969:8). Thus, CPs can be regarded as indices or as "Exponenten von nach inhaltlichen Merkmalen geschiedenen Nominalklassen" (Kolver 1979:1); they represent the semantic (sub) structures of a (classifier) language (see Friedrich 1970:379). The critical questions to be answered are, What are the semantic criteria and principles this kind of classification is based on? and, Are the classifications in different languages culturally determined? Before I attempt to answer these questions, I want to emphasize that the classificatory systems of the various numeral classifier languages are usually not comparable to folk taxonomies, but must, more often than not, be regarded as paradigms (see Becker 1975:111; Berlin 1968; Berlin et al 1973; Burling 1965; Conklin 1962:124; Ellen 1979:7-9; Foucault 1966:108, 110; Frake 1969:34; Haas 1942; Barz and Diller 1985:176; Hundius and Kolver 1983:204; Miram 1983; Saul 1965; G. Senft 1987b; Tyler 1969:7-9). In taxonomies, the respective nominal referents are classified on the basis of the objectively perceptible and verifiable features. In paradigms, the single nominal referents are categorized in contrastive relation to other nominal referents. Mixed forms of taxonomic and paradigmatic classification do exist, but they are exceptional; moreover, such mixed forms of classification depend on the inventory of CPs these languages display. The inventory of CPs in classifier languages ranges from 2 to 528 (or even 730) (see Adams and Conklin 1973:9; Berlin 1968; Miram 1983:103). Descriptions of the criteria that structure classifying systems generally give the following features: + human; human and social status; human and kinship; + animate; sex; shape/dimension; size; consistency; function; arrangement; habitat; number/amount/mass/group; measure; weight; time; action; + visible (see Adams 1989; Adams and Conklin 1973; Allan 1977; Becker 1975; Benton 1968; Burling 1965; Denny 1979; Friedrich 1970; Haas (continued from page 7}
(1968) also describes "action classifiers" in Tzeltal; that Harweg (1987b) uses the term Zahleinheitsworter as a synonym for Numerativ; that Hoa (1957:128) introduces the term "semelfactive classifier" for a "type of classifier which indicates single action", referring to certain numeratives in Vietnamese; that Adams (1989):177, 182) describes "the general or cannibalising classifier"; that Hiranburana (1979:39-40) uses the terms "reduplicative classifier" and "imitative classifier"; that Fischer (1972:69, 77) speaks of "isolierte Klassifikatoren" (i.e., one classifier classifies only one noun) and mentions "metric classifiers" (i.e., classifiers for numerical and temporal units); and that Craig (1992:285-286) mentions "genitive classifiers" that are also called "relational" or "possessive classifiers". See also Seller (1986).
10
Classificatory Particles in Kilivila
1942; Hiranburana 1979; Hoa 1957; Kaden 1964; Miram 1983; Royen 1929:82-83, 125-137, 142, 256, 396-397).12 Classificatory systems are usually described by feature lists that give the features in a relatively free order; however, there have been a few attempts to order the features hierarchically (see, e.g., Goral 1978:38; cf. Craig 1986:5-6). Becker (1975) describes the Burmese system of CPs hierarchically in the form of concentric circles; Miram (1983) proposes functional diagrams—Flussdiagramme—to describe the system of Yucatecan Maya (I will discuss this approach in more detail in section 1.4.1). What must be emphasized here is the fact that most, if not all, of these features represent semantic categories that are fundamental in all languages. For example, Friedrich (1970:404) characterizes the feature "shape" as the "ultimate semantic primitive" (see also Allan 1977:302). Moreover, it should also be noted that—at first sight, at least—these features seem to be universal (Lyons 1977b:466). However, a closer look at the CPs that constitute the semantic domains in individual languages on the basis of these features reveals that these general, probably universal, categories are defined in culture-specific ways (Berlin 1968:35). It is also evident that the boundaries between the individual semantic domains are rather fluid (Rosch 1978:36, 1977:4, 15, 18, 21). Thus, Craig (1986:1), on the basis of prototype theory, claims rightly that ". . . categories . . . should be described as having fuzzy edges and graded membership . . ." (see also Posner 1986; Givon 1986).13 Therefore, the description of semantic domains within any numeral classifier language requires a sound analysis of how these domains are constituted, that is, which features are relevant for the definition of the semantic domain (see Lenneberg 1953:468; Rosch 1978:28). This ethnosemantic descriptive and analytical research is rather complex and presupposes the linguist's thorough knowledge of the language described. This may explain why, at least in my opinion, typological, comparative studies on classifier languages, which necessarily use a general approach, neglect this microlevel of ethnosemantic analysis.14 In these studies, relating the general results of the description of semantic domains to a specific language is possible only in a rather indirect, intermediate way. Moreover, it may well be that the complexity of ethnosemantic analysis is the reason that most descriptions of Classificatory systems thrust the equally important morphological/syntactic role of CPs into the background. Now, what about the actual use of the CPs that constitute such complex 12
Apparently no classifier language makes use of a color classifier (see Adams 1989:5; Allan 1977:297; Asmah 1972:94; Berlin and Romney 1964:80; Carroll and Casagrande 1958:27-29, 31; Lee 1987:397; but see also Royen 1929:151). 13 With respect to prototypes, see Rosch (1988); see also Mufwene (1980b:30, 31, 33, 36; 1980c;1983) who describes prototypes as "underlying principles of cognitive classification" and discusses possible ambiguities because of their "metaphorical extension(s)". For a critical discussion, see MacLaury (1991); see also Fischer (1972:68), who pleads for a historical foundation for the structure of so-called semantische Felder. 14 Here the eminent exception is Royen (1929)!
Introduction
11
systems?15 As mentioned previously, the inventory of CPs in classifier languages varies from 2 to 730 lexemes. In languages with many CPs, it must be emphasized that some CPs can refer to any nominal denotatum. That is, even with inherent classification, CPs can be used to specify special aspects of relatively general nominal concepts. In other words, complex systems of CPs allow the possibility of referring to a noun within its semantic domain either by the general, characteristic, "unmarked" CP or by a more specific CP.16 The choice of the appropriate CP occurs on the semantic level. It can be independent of the intended speech act, and thus attains stylistic denotation, meaning, and significance (see Becker 1975:113; Burling 1965:259; Goral 1978:26). Individual speakers use these options in their choice of CPs.17 They may even be innovative by using a certain CP in a metaphoric way. Some linguists (e.g., Becker (1975:113), in his work on Burmese CPs) even claim that the actual "use of classifiers . . . is in part an art" (see also Rosch 1977:42). This implies that analyses of CP systems must take into account the function of CPs as sociolinguistic variables (Labov 1972:237; see also Hla Pe 1965:170). Thus, we can conclude that all CPs do have meaning (Allan 1977:290; see also Adams 1989:7-8, 192; Adams et al. 1975:14, 17; Berlin and Romney 1964:79, 82; Mufwene 1984: 201, 203-204, 214; Schafer 1948:412). The problem of how to describe this meaning leads us to basic linguistic methodological considerations. 1.4 METHODS AND AIMS OF TRADITIONAL CLASSIFIER STUDIES— A BRIEF SURVEY
This section discusses studies on classifier languages in general, as well as studies on Kilivila in particular. 1.4.1 Studies on Classifier Languages The literature cited so far includes typologically oriented comparative studies, on the one hand, and empirical descriptive studies of individual languages, on the other. A discussion of the methods that typological comparative studies use is unnecessary here. Rather, methodological techniques used in 15 With the exception of Downing (1986) and Erbaugh (1986) this question is not answered in any of the literature cited here. 16 Benton (1968:111) states, "The 'characteristic' or unmarked meaning of a form is generally that which is most likely to be encountered in a particular context and may not have any direct connection with a specific set of inherent features", and (Benton 1968:137), "Within a domain, a change in classifier generally signals a change of meaning. That is, certain semantic features of the noun . . . are highlighted by a particular classifier . . . ." See also Adams (1986:241-243) and Lehmann (1979:171), who mentions ". . . the semigrammatical (e.g., sarcastic or insulting) deliberate choice of a 'wrong' classifier . . . ." See also Allan (1977). "With respect to the topic "frequency of usage", see for example, Rosch (1977:38). See also Mufwene (1980a) who notes the discrepancy between CP inventory and actual CP usage.
12
Classificatory Particles in Kilivila
empirical descriptive studies of individual languages will be described, based on Berlin's (1968) and Miram's (1983) monographs; other studies will then be discussed in light of the questions raised by these monographs. In his work on the system of CPs in the Mayan language Tzeltal, Berlin presented two literate, Spanish-speaking, native Tzeltal speakers with 4,410 "phonemically possible forms". Using these forms, he elicited from the two consultants 528 numeral classifiers, which were counterchecked with the help of three more consultants in "informal and nonsystematic checks of reliability" (Berlin 1968:13, 19-20; see also Berlin and Romney 1964:80-81). The elicited CPs are described according to their features. Using the same consultants and a questionnaire, Berlin ordered the CPs hierarchically and according to the semantic domains they constituted. Thus, the Tzeltal CPs were described on the basis of the judgments of only two—or, if we include the three counterchecking Tzeltal speakers, five—consultants. In her study on the CPs in Yucatecan Maya, Miram (1983) also worked with a literate and bilingual consultant, Edilberto Ucan Ek; he became her assistant and also did transcriptions for her. Miram's data are based on observations of ordinary language spoken in a natural context, as well as on interviews with consultants. During these interviews, she (or her assistant) checked lists of possible or already described CPs and elicited further, "new" CPs. It is unclear, however, which kind of data she is referring to in her study. Miram elicited 730 CPs; on the basis of clearly defined criteria, she reduced these CPs to 225 forms that were semantically analyzed. Her assistant grouped 194 of these CPs into 47 semantic domains. What is decisively new in her approach is her use of functional diagrams based on these data. The diagrams are intended to simulate a native speaker's decision processes in choosing a certain CP. This form of data presentation allows a description of the semantic domains as "open" systems that may be linked with each other in various ways (see Ellen 1979:17). The value of Miram's novel approach is unfortunately partially diminished by the somewhat unclear description of her consultants' contributions and by the important role her assistant played in the processes of data collection and data processing. Linguists experienced in field research will probably agree that Miram's assistant falls into the category of the so-called marginal native (Freilich 1970). This is true of Berlin's consultants, too. In general, all data gathered with the help of such consultants must be looked upon with a certain amount of skepticism. Although I am critical of the databases used in both of these studies because they resulted mainly from work with only a few consultants, it should be pointed out that Berlin and Miram at least reveal the methods of their data collection. Although there are a few exceptions (e.g., Carpenter 1991), many other investigators either do not mention the methods used in data collection or claim that theirs is a comparative study, which need not use empirical data gathered by the author. Some studies are based on data presented in already published papers and on judgments of native speakers and (foreign) trained linguists. Kolver's excellent study on Central Thai is an example (Kolver 1979; see also Unterbeck 1990b:3-4, 61). Nevertheless,
Introduction
13
in general descriptive studies of the CP systems of individual classifier languages have a number of significant methodological shortcomings. The research interests of published CP studies can be summarized as follows: 1. Descriptions of the syntactic role of CPs. 2. Descriptions of the semantics of CPs. 3. Descriptions of CPs with respect to their constitutive function for semantic domains, as well as lists of these semantic domains. 4. Comparisons of semantic domains constituted by CPs in different classifier languages, with respect to the question of language universals. To my knowledge, only nine empirically founded studies deal with the question of first-language acquisition of CP systems (Carpenter 1986, 1991; Chin 1989; Erbaugh 1986; Gandour et al. 1984; Luke and Harrison 1986; Matsumoto 1985; Sanches 1977); Clark (1976) unfortunately indulges in sheer speculation.18 However, the studies of Allan (1977:295), Burling (1965:261), Goral (1978:4), Haas (1942:201-202), and Tversky (1986:72-73) provide some hints toward possible answers to the question of how the CP system is acquired by native speakers. With respect to data showing CPs in actual speech, the currently available CP studies—with the exception of Becker 1986, Downing 1986, and Erbaugh 1986—only present isolated sentences as examples to support the syntactic and/or semantic description of the CP systems. Context influence (Rosch 1978:42; Strawson 1950:334; cf. Levinson 1983:172-177) on the speaker's choice of a certain CP is a problem sometimes mentioned but not illustrated by relevant data. Symptomatic of the state of the art with respect to this question is Berlin's (1968:23) statement (which he does not indicate is a hypothesis) that "The actual occurrence of numeral classifiers in recorded texts is relatively low . . . ." He then concludes that the occurrence of CPs in various contexts can be ignored in analyzing the CP systems of individual languages, and supports this conclusion with the statement that ". . . only systematic elicitation for classifiers per se . . . can hope to claim partial exhaustiveness" (see also Craig 1986:8). Thus, Berlin supports Burling's (1965:264) guiding methodological principle for the analysis of CP systems: Seeing the problems which arise in the attempts to bring order into the set of classifiers, one may feel that the best available "analysis" so far is simply the list of classifiers with their definition. 18 For the acquisition of Niger-Congo noun class and agreement systems, see Demuth et al. (1986:463-465). Noun classes and agreement in Sesotho acquisition are discussed in Demuth (1988) and Demuth (1992). For a discussion of how and when number and countability distinctions are acquired in English, see Mufwene (1984:207-208). This criticism and some of the other statements presented here do not hold for studies on the ASL classifiers; see, for example, Kantor (1980). For a criticism of Clark (1976) see Goral (1978:4), also Chin (1989:156-158).
14
Classificatory Particles in Kilivila
For the lexicology of a classifier language, this is an important conclusion, of course; however, for a CP analysis that tries to penetrate into the microlevels of the CP system, the static listing of CPs is certainly not entirely satisfactory.19 Finally, there is virtually no information about processes of language change in progress affecting CP systems.20 1.4.2 Studies on Kilivila The previous section surveyed rather briefly publications on numeral classifier languages other than Kilivila, the Austronesian language whose CP system is the focus of the analyses presented here.21 Milner (1963:66) poses the following question: Is an Oceanic linguist to concentrate on one language at a time to treat it with the same searching analysis that is now taken for granted in the study of, say, Latin or French, at the risk of too great a degree of specialisation, or is he to be content with general surveys and comparative work founded generally on second-hand and superficial knowledge of his material? Faced with these alternatives, (see also Vachek 1976:225-226), I decided on the first course, the analysis of the CP system of one language, namely, Kilivila. Kilivila CPs are first mentioned by Fellows (1901) and Ray (1907). Baldwin (no date), who worked as a missionary on Kiriwina Island, lists 75 classifiers in his unpublished fragment of Kilivila grammar, translates them into English, and describes aspects of their morphological role and function. Malinowski (1920) describes 42 CPs with respect to their morphological, syntactic, and semantic functions. Capell (1969:61) lists 44 CPs in his description of noun class division in Kiriwina.22 Capell (1971:273-274) and Lithgow (1976:461, 465-467, 480, 488-490) also refer briefly to the Kilivila CP system. Ralph Lawton's unpublished Master's thesis "The Kiriwinan Classifiers" (Lawton 1980) is the most comprehensive description of Kilivila CPs so far, and thus warrants a brief review. After a description of the morphological 19 That it is possible to describe the semantic domains constituted by the CPs as dynamic, open systems is convincingly demonstrated by Miram's (1983) monograph. 20 As far as I know, only Asmah (1972:96); Becker (1975:120); Demuth et al. (1986); Dixon (1986:110-111); Erbaugh (1986); Jacob (1965); and Kolver (1982a:107, 115, 120) discuss possible processes of language change affecting CP systems. 21 The CP system of Kilivila is mentioned briefly in the following publications: Allan (1977:285, 288, 290, 295-296, 299, 300, 303-305); Chin (1989:24-25); Friedrich (1970:401-403); Greenberg (1975:18-19; 1978:78). 22
However, the CP ukdu that Capell mentions is not (or no longer?) known to Kilivila native speakers; one explanation for this may be that this CP does not agree with Kilivila syllable patterns (see G. Senft 1986:20-22).
Introduction
15
role the CPs fulfill and a discussion of the special relationship between CPs and nouns, Lawton presents a semantic description of 147 CPs. He divides these into two groups: Group 1 comprises 34 CPs, which "specify whole items in terms of their features or properties" (Lawton 1980:80). Group 2 comprises 113 CPs, "which classify items in terms of some modification they have undergone. Modification of items is conveniently divisible into three categories labelled activity, partition and arrangement" (Lawton 1980:81). Lawton then subdivides the CPs of group 1 into the following three classes: 1. Basic property specifiers which are subclassified according to the features + animate, + human, dimension, and residue. 2. Subclassifying CPs within the semantic domains constituted by the basic property specifiers. 3. Residue. The individual CPs are listed and described, and a phrase or sentence is given as a reference for the use of the respective CP. In his semantic description of the CPs of group 2, Lawton subclassifies the partitive classifiers according to the features "topographical, parts within wholes, pieces, multiple reference", and the arrangement classifiers according to the features "inherent/non-inherent arrangement"; the feature non-inherent arrangement is itself further subcategorized according to the features "distributional", "configurational", and "quantitative". The individual CPs of group 2 are presented in the same way as the CPs of group 1. Lawton then discusses the CPs in connection with adjectives, verbal expressions, associations, and metaphors; the appendix presents a list of the 147 CPs according to Lawton's classification principles and in alphabetical order. Some of Lawton's expositions are problematic, but the whole topic is complex. Thus, I shall not criticize Lawton (1980) in detail here. However, I want to note that almost all the points of criticism mentioned in the discussion of the studies of classifier languages in general hold for Lawton's thesis as well.
Chapter
2
Classificatory Particles in Kilivila: Grammatical and Discourse Functions In this chapter, we will look at the grammatical and discourse functions1 of the Kilivila system of CPs, including the morphological relevance of CPs with respect to Kilivila inflectional morphology and the functions that are assigned to CPs in Kilivila. 2.1 MORPHOLOGICAL RELEVANCE
The Kilivila system of CPs contains at least 177 formatives. (Section 3.3.1 lists these CPs in detail.) I assume that with all the subtle and very specific differentiations possible, there are probably more than 200 CPs in Kilivila. Moreover, if we keep in mind all the pragmatic functions CPs can serve, the Kilivila CP system can even be regarded as basically open; here it has to be noted, however, that at least so far no loanword has been incorporated completely or in part into the inventory of Kilivila CPs (see G. Senft 199la, 1992a). The system of noun classification in Kilivila is an important means of word formation with (1) all but one of the demonstrative pronouns, (2) one 1 Kilivila has a fourfold series of possessive pronouns, partly realized as free pronominal forms, partly realized as pronominal affixes. One of these series is produced in a specific semantic context, referring to food only; the other three series are used to distinguish different degrees of possession: one series marks inalienable possession and two series mark nonedible alienable possession (G. Senft 1986:47-54). These possessive pronominal forms classify the Kilivila noun, too, of course (see also Royen 1929:186; Wurm 1981). However, I do not deal with this kind of nominal classification in this book, but rather present research on the specific system of formatives that consists of quantifiers, repeaters, and noun classifiers proper, which all are referred to with the general term—classificatory particles—that Malinowski coined for them. Finally, I am sure that Kilivila has verbal derivational classificatory prefixes similar to those described by Ezard (1978) for languages of the Massim cluster; however, as stated elsewhere (G. Senft 1986:28), I have omitted all consideration of derivational morphology in my description of Kilivila.
16
Grammatical and Discourse Functions
17
form of (numerical) interrogative pronouns/adverbs, (3) two classes of adjectives, and (4) numerals. These word classes require concord with the class of the noun to which they refer. This concord is secured by the CPs, which are infixed or prefixed to the respective word frame or word stem. I have described these processes of word formation and the syntactic aspects of constituents with CPs in detail elsewhere (G. Senft 1985a:374-379, 1986); however, I will give a rather general account of the processes of word formation here. With the exception of the purely exophoric demonstrative pronoun besa or beya 'this' (with a deictic gesture), all other demonstrative pronouns consist of a fixed morphological frame, formed by the word-initial morpheme ma- (or, according to phonological rules, m- or mi-), the word-final morpheme -na; and an infixed morpheme, which is the CP. To distinguish between singular and plural, there is also a plural marking morpheme -si-, which is infixed between the CP and the word-final morpheme -na. Demonstrative pronouns formed in this way express the concept of 'this/these'. To express the deictic concept of 'that/those', the morpheme -we- is infixed in singular forms between the CP and word-final -na and in plural forms between the plural marker -si- and word-final -na. To express the kind of deictic concept that comes close to the English demonstrative 'yonder', the Kilivila speaker takes the forms of the demonstrative pronouns expressing the concept of 'that/those' and changes the final vowel /a/ of the word-final morpheme -na to an /e/ that is lengthened and has a minor accent. There are three classes of adjectives in Kilivila. One class must be used without CPs, the second may be used with or without CPs, and the third must always be used with CPs that are prefixed to the word stem. The numerals or, more precisely, the cardinal numbers in Kilivila consist of the word stem and a prefixed CP. There is also one form of an interrogative adjective or adverb that consists of the word stem -vila and a prefixed CP (for CPs with an interrogative quantifier see de Leon 1988:65-66). I conclude this brief account of the processes of word formation in Kilivila with the presentation of two sentences containing all four word classes involved in the system of noun classification (see G. Senft 1989): (1) Kevila waga lekotasi? ke-vila waga le-kota-si? wooden-how many canoe 3Ps.Past-arrive-Plural2 'How many canoes arrived?' 2 I tried to use as few abbreviations as possible for glosses and in the Kilivila orthography (see Senft 1986). Here is the list of abbreviations used for these purposes:
Ps PP-IV '
person fourth series of possessive pronouns (see note 1 to this chapter) glottal stop
18
Classificatory Particles in Kilivila
(2) Keyu waga makesina kemanabweta (lekotasi). ke-yu waga ma-ke-si-na wooden-two canoe this-wooden-Plural-this ke-manabweta (le-kota-si). wooden-beautiful (SPs.Past-arrive-Plural). These two beautiful canoes (arrived).' The speakers of these sentences in referring to canoes must indicate the noun class of 'canoe' with the CP for 'wooden things'—ke—in the interrogative pronoun, in the numeral, in the demonstrative pronoun, and in the adjective. Concerning the morphological relevance of the CPs in Kilivila, I have already mentioned one function of these formatives, namely, to secure concord between the noun and the four word classes involved. The next section describes the functions of CPs in Kilivila in more detail.
2.2 FUNCTIONS OF CPS IN KILIVILA I have already mentioned some functions CPs fulfill in classifier languages in general. I will now describe the functions of CPs in Kilivila. I will illustrate these functions by isolated phrases or sentences (see G. Senft 1989); however, I refer anyone interested in the actual realization of CPs in Kilivila speech production to the already published parts of my Kilivila text corpus (Eibl-Eibesfeldt et al. 1987; B. Senft and G. Senft 1986; G. Senft 1985c-f, 1986, 1987c,d, 1991b,d, 1992b).
2.2.1 Referential Function—Concord As mentioned previously, CPs perform the referential function of securing concord between the nouns and the word classes that use CPs in their word formation. This concord, of course, creates redundancy in the information conveyed by a sentence, as was illustrated in examples (1) and (2). The reference of the various word classes is unequivocal, and the redundancy is obvious: Trobriand canoes are made of timber; they are "wooden things" (this aspect of redundant information will be discussed in section 2.2.3). The complex inventory of CPs allows speakers to classify a noun "temporarily" (Berlin 1968:175), that is, to emphasize certain characteristics of the referent of the noun. This is illustrated by the following examples (from G. Senft 1985a:380-387): (3) natala yena na-tala yena animal-one fish 'one fish'
Grammatical and Discourse Functions
19
(4) kevalalima yena kevala-lima yena batch drying-five fish 'five batches of smoked fish' (5) oylalima yena oyla-lima yena string-five fish 'five strings with stringed-on fish' (6) makupona yena ma-kupo-na yena this-two string-this fish 'these two strings of fish' (7) mapwasasina yena ma-pwasa-si-na this-rotten-Plural-this 'these rotten fish'
yena fish
These examples first represent the CP na meaning 'animals' and then illustrate a set of the noun-modifying group of CPs that specify the noun with respect to quantity, order, arrangement, and condition or state. Example 8 presents the two gender-specifying CPs to/te and na (now meaning 'persons of female gender') and the age-subclassifying CP gudi. (8) Bibodi tetala natala guditala bi-bodi te-tala na-tala gudi-tala 3Ps.Fut.-benefit male-one female-one child-one 'It will benefit each man, woman, and child.' The noun phrase in example (9) (see Lawton 1980:49) illustrates the semantic power of CPs. (9) kai mabubosina kwelatolu kai ma-bubo-si-na kwela-tolu wood this-cut across-Plural-this potlike-three 'these three potlike sawed-off sections of timber' Example (10) shows that CPs can also be used metaphorically. Here, a speaker refers to a 'dinghy' as a 'child-canoe': (10) Kugisi magudina waga kekekita okopo'ula waga dimdiml ku-gisi ma-gudi-na waga ke-kekita 2Ps.-look this-child-this canoe wooden-small okopo'ula waga dimdim behind canoe white man 'Look at this small dinghy behind the motorboat!'
20
Classificatory Particles in Kilivila
All these examples illustrate the referential function of CPs and their semantic power. A closer look at some of these examples shows other (grammatical) functions that CPs perform, as described in the next section. 2.2.2 Nominalization, Plural Marking, Numeralization, and Verblike Expressive Functions In example (8), the numerals tetala, natala, guditala are translated as nominal expressions. This is legitimate, especially if we assume that the nouns of the three noun phrases (tetala tau 'one man', natala vivila 'one woman', and guditala gwadi 'one child') were omitted. This analysis—which is possible because of redundancy conveyed by CPs—assigns proper nominal status to the numerals. We also find this kind of nominalization with demonstrative pronouns and adjectives (see G. Senft 1985a:384). The noun phrases in (11) and (12), as well as those in (4)-(6), illustrate the plural-marking function of CPs. (11) makena nuya bwaveaka ma-ke-na nuya this-wooden-this coconut 'this big coconut tree'
bwa-veaka tree-big
(12) mapo'ulana nuya bwaveaka ma-po'ula-na nuya bwa-veaka this-plantation-this coconut tree-big 'this plantation of big coconut trees' Besides plural marking, we also find some CPs that quantify numeralization, a function independent of that of numerals proper. The noun phrase in example (6), given previously, illustrates this function. In examples (4), (5), (7), and (9), it is clear that some CPs take over the function of verblike expressions within a noun phrase. This is especially true for CPs that specify or refer to certain activities (see G. Senft 1985a:385). So far we have dealt with CPs only on a sentence or phrase level. In the next section we look at CPs as they are used in actual discourse. 2.2.3 Redundancy, Ellipsis, and Discourse Coherence Example (8) shows that noun phrases can be composed of numerals without the respective nouns to which they refer. I explained this type of noun phrase construction by proposing that the noun is omitted and that the other word class (in this case, the numeral) acquires nominal status. Malinowski (1920:59-60) hinted at such an interpretation and compared this type of sentence with elliptical utterances in English. Sentences that are constructed like example (8) are, indeed, quite frequently produced in Trobriand discourse. Trobriand Islanders introduce a certain nominal
Grammatical and Discourse Functions
21
denotatum explicitly. If they then want to further refer to this noun in discourse, using numerals, demonstrative pronouns, and adjectives, they usually omit the noun. This is only possible because the CPs represent the omitted nouns in a quasi-fragmentary way and the anaphoric reference of CPs secures semantic concord beyond sentence boundaries. Now we can explain why we sometimes find redundant information within the noun phrase: the information redundancy given by the CPs within a Kilivila noun phrase enables the omission of the noun without any loss of information, even beyond sentence boundaries. Thus, CPs perform the important function of securing coherence in discourse. As a general rule, a noun can be elided as long as it is not reclassified (e.g., for stylistic reasons) by another CP. If this occurs, the noun must be overtly realized again as a constituent of the noun phrase to ensure unequivocal and unambiguous reference. In my sample of transcribed Kilivila speech data, I have one (rather extreme) example where a speaker (Tomalala, Consultant No. V16) introduces a nominal referent to which he then refers 16 sentences (78 words, 113 morphemes) later with the appropriate CP; nevertheless, the reference is unequivocal. Examples (13)-(15) illustrate these functions of CPs. (13) Atatai tataba. Tauwau Tabalu mtosina makena si koni. a-tatai tataba tauwau tabalu 1Ps.-carve tataba-board men Tabalu-subclan m-to-si-na ma-ke-na si koni this-male-Plural-this this-wooden-this their sign of honor 'I carve a tataba-board. These men belonging to the Tabalu-subclan, this is their sign of honor.' Here, the speaker refers to a certain board with carved patterns that marks houses, food houses, and canoes as the personal property of men belonging to the Tabalu-subclan. The reference of the two demonstrative pronouns is unequivocal because in this context the CP to can only refer to the noun tauwau, and the CP ke can only refer to the noun tataba. (14) Tauwau pela emesi bilebusi. Ekokwa'usi kebila mabudanaga ekugwasi emesi. tauwau pela e-me-si bi-lebu-si men for 3Ps.-come-Plural 3Ps.Fut.-take-Plural e-kokwa'u-si 3Ps.-weave-Plural
kebila stretcher
ma-buda-na-ga this-group-this-Emphasis
e-kugwa-si e-me-si 3Ps.-first-Plural 3Ps.-come-Plural The men have come to take him with them. They have woven a stretcher, the men belonging to this group who were the first to arrive.'
Classificatory Particles in Kilivila
22
Here, the speaker uses the CP buda with the demonstrative pronoun in the second sentence to refer unequivocally to the noun tauwau in the first sentence (see G. Senft 1985d:481). (15) 0 davalusi esisusi tommota topaisewa. Vivila nasalau, tauwau tobugubagula Tommota gala todubakasala, kena kumwedona enukwalisi bubunesi bwena. o da-valu-si e-sisu-si tommota in 1Ps.incl-village-Plural 3Ps.-live-Plural people to-paisewa human beings-work
vivila woman
na-salau female-busy
tauwau men
to-bugubagula tommota gala to-dubakasala male-work in the garden people not human beings-rude kena kumwedona e-nukwali-si bubune-si bwena but all 3Ps.-know-Plural manners-their good 'In our village live people taking pleasure in their work. The women are busy, the men are good gardeners. The people are not rude, but all have good manners.' Example (15) illustrates that, in general, reclassification of a noun does not allow it to be omitted. To emphasize the different characterization of men and women, on the one hand, and all villagers, on the other, the nouns can hardly be omitted. The speaker uses the CP to to refer to 'human beings' and to 'persons of male sex'. The CP na is used to refer to 'persons of female sex'. If the speaker did not use the noun tommota in the last sentence again, then this sentence would refer to 'persons of male sex' only (see G. Senft 1985a:387-388).
2.3 SUMMARY The grammatical and discourse functions CPs perform in Kilivila are following (see G. Senft 1991c): 1. They play an important role in the word formation of all numerals, all demonstrative pronouns (with the exception of the general demonstrative besa), some adjectives (see G. Senft 1986:85-88), and one interrogative adverb or numerical interrogative pronoun. 2. They mark concord between nouns classified and the word classes containing the CP. 3. They classify and specify their nominal referents, inherently as well as temporarily, in many different ways and with much semantic power. 4. They can nominalize all numerals, some adjectives, and all demonstrative pronouns (with the exception of besa).
Grammatical and Discourse Functions
23
5. Being collective terms (Kollektiva, see Royen 1929:595, 597, 601, 612, 251, 512), they can mark plural on the nouns to which they refer (see also Adams and Conklin 1973:8-9). 6. Some can perform verblike functions within noun phrases in a sentence. 7. With their anaphoric referential function, they can constitute noun phrases that are comparable to elliptical utterances: once a noun has been introduced, as long as it is not reclassified, the following references to this nominal denotatum may consist of numerals, adjectives, and/or demonstrative pronouns only, that is, the noun itself is then no longer realized; it is ellipsed in the noun phrases (see also Adams et al. 1974:1, 10-12; Buhler 1978:155-158). 8. With their anaphoric referential potential, they can perform the function of preserving coherence in discourse (see also Hopper 1986).
Chapter
3 The System of Classificatory Particles in Kilivila In this chapter, an analysis of the Kilivila CP system is provided. Section 3.1 describes the research aims and methods; section 3.2 presents the data collected, as well as a first interpretation of the data; and section 3.3 presents the results of my research, answering the questions raised in section 3.1. The tables presented in this chapter are numbered according to the following logic: Tables 3.1.1 and 3.1.2 present the consultants and the questionnaire used to elicit CPs. Tables 3.2.0-3.2.216 present the data on which the analyses are based. Tables 3.3.1-3.3.7 present ordered data for, and results of, analyses with respect to the inventory, the acquisition, and the production of CPs. Tables 3.4.1-3.4.22 present ordered data for, and results of, the CP semantics as well as the semantic domains constituted by the CPs. The table numbers do not correspond to numbered sections and subsections within this chapter. 3.1 AIMS AND METHODS In carrying out linguistic research in the Trobriand Islands, one of my main interests—stimulated by reading Malinowski's 1920 article—was to work on the system of classificatory particles, as soon as I was able to master the language in an appropriate way (this is, in my opinion, a necessary prerequisite for performing such a study). I wanted to answer the following questions: 1. Number of CPs. a. Which formatives constitute the Kilivila CP system? Malinowski (1920:44)discusses42CPs,Capell(1969:61)lists44CPs,and Lithgow (1976:480) mentions more than 50 noun classes. b. What is the actual occurrence of CPs in recorded texts? 24
The System of Classificatory Particles in Kilivila
25
2. Acquisition. a. What is the general temporal progress of the CP acquisition process? b. Are there any differences in the CP acquisition process with respect to the production of demonstrative pronouns, adjectives, and numerals that use CPs as morphemes in their word formation? c. Are there any gender-specific differences in the CP acquisition process? d. What is the order in which the individual CPs are acquired, and why are the individual CPs acquired in this order? e. Are there any parallels between Kilivila CP acquisition and CP acquisition data for other classifier languages? f. Are there any parallels between the Kilivila data and acquisition data with respect to demonstrative pronouns, adjectives, and numerals for such Indo-European languages as English and German? 3. Realization of CP types in actual speech. a. Can a classification of the CP system be devised? b. How do consultants realize the individual CP types, and can we observe any processes of language change in progress? If I had an opportunity to observe language change, I wanted to answer the following questions: i. Why are CP types affected by processes of language change, and who opposes and who fosters language change? ii. What are the possible consequences of language change for the grammar of Kilivila? 4. Semantic domains. a. What semantic domains are constituted by the Kilivila CP system? b. What are the dynamics of the semantic domains constituted by the CPs and what rules do speakers use in their production of a certain CP? This was of special interest to me because I assume that a complete description of a classifier system must explain why a speaker produces a certain CP to refer to a certain nominal concept. Thus, the linguist's description should predict which CPs a speaker will produce (and with what relative frequency) to refer to certain nominal concepts. After answering these questions, my goal was to discuss possible interdependences between language, thought, and culture (see chapter 5). To achieve these aims, some basic methodological decisions had to be made, especially concerning the method of data gathering. My overall corpus of recorded and transcribed Kilivila speech data had to be checked with respect to the CP types produced when I returned from my fieldwork. However, I also wanted to do a study in the field on the system of CPs. Thus, I started my first informal pretests on how to elicit CP data in January 1983, after having lived in Tauwema village for five months. Of course, my corpus then was not as extensive as it was at the end of my fifteen months
Classificatory Particles in Kilivila
26
of field research (see: G. Senft 1985b:131-134); therefore, it could only serve an auxiliary function with respect to this specific research interest. My first aim was to make an inventory of Kilivila CPs. Using Malinowski's (1920) list, and Baldwin's list in his unpublished (and undated) manuscript "Biga Boyowa" (see G. Senft 1987a:102, 119), as well as writing down all the CPs I heard in everyday conversation in the Trobriand Islands and checking my transcriptions, I arrived at a total of seventy formatives. I checked these with my main consultants and then compiled a list of CPs to be used as a basis for further data gathering with other consultants. At this point, a decision had to be made with respect to consultants: I could have worked either with consultants living in different villages scattered all over the Trobriand Islands or with consultants living in Tauwema village, my place of residence during my field research. I chose the latter alternative because of transportation problems and because working with people in various locations would have meant a serious neglect of other interests I wanted to pursue during my field research. Thus, I conducted my research on CPs using the language variety of Kilivila that is spoken by the inhabitants of Tauwema village on Kaile'una Island (see G. Senft 1986:6-8). At the time of my research, Tauwema had 239 inhabitants (120 adults: 58 women, 62 men; 119 children; 52 girls, 67 boys). I decided on a random sample of 60 consultants (30 female and 30 male), representing five different age groups ranging from approximately 4 to 75 years. Table 3.1.1 shows the age and gender of the consultants with whom I worked (appendix A presents the individual consultants with their consultant number, name, age, and clan). This sample included 20% of the children and 30% of the adults living in Tauwema. Because there is no official or reliable registration of dates of birth on the Trobriand Islands, the age of all consultants had to be estimated. Pretesting had shown that there was no point in trying to elicit CPs from children younger than 4 years of age; children start to acquire these formatives around the age of 4 ( + 6 months). Within all five age groups, I tried to balance the consultants according to gender and age, especially in the two adult groups that spanned ages 14 to 39 years (in the actual test: 11 to 38 years; see appendix A). As demonstrated elsewhere (B. Senft and G. Senft 1986, Senft 1987b), the size of this sample is adequate to meet all the requirements that the aims of studies like the one presented here demand. Table 3.1.1. Consultants
Age group
I. II. III. IV. V.
Children, approximately 4-7 years Schoolchildren, approximately 8-14 years Adolescents, approximately 15-20 years Adults, approximately 21-35 years Adults, approximately 36-75 years Total
Female
Male
All
6
6 6 6 6
12 12
6 6 6 6
30
12
6
12 12
30
60
The System of Classificatory Particles in Kilivila
27
Regarding an appropriate method of data gathering, I decided to compile a questionnaire to elicit data on CPs. The questionnaire used the 70 CPs in my inventory. To elicit these formatives, I asked my main consultants in various sessions for typical nouns or noun phrases that are classified most appropriately and unequivocally by the respective CP and thus would elicit the production of this CP. These typical nouns and noun phrases, the qualities they described, and the class indicated by the particular CP were noted. The stimuli for the CPs elicited by the questionnaire were presented in random order. Table 3.1.2 presents the questionnaire with the CPs in alphabetical order, together with the then known meaning(s) of the respective CPs (for all meanings of the CP types found in this study, see Table 3.3.1 in section 3.3.1) and with the eliciting nouns or noun phrases and their translation. The actual order in which the CPs were elicited in the test is indicated by a number (in the third column in Table 3.1.2). Table 3.1.2. The questionnaire No. in test
CP to be elicited
Eliciting noun/noun phrase
bobu 'block cut off bogi 'night'
bobu 'log, round timber' bogi 'night'
65 45
bubwa '(parts) cut off buda 'group, team'
babwabutou kai 'I will cut off wood' boda 'group, team'
20
bukwa 'fruit cluster' bwa 'trees, wooden things' deli 'group, company on the move'
budubadu nuya 'many coconuts' kai 'wood, tree' deli 'group'
17 44
dull 'cluster, bundle'
duli 'cluster, bunch, bundle' duyava 'door, entrance, window'
duya 'door, entrance' gili 'row' giwi 'cut' gudi 'child' gula 'heap, group' gum 'bit, small piece' iga 'name' kabisi 'compartment of a foodhouse' kabulol 'protuberances, point' kabulo2, kabulu 'village sector' kada 'road, track' kaduyo 'door, entrance' kai 'stone blade'
vakala 'belt of spondylus shell discs' giwi 'cut'
55
56 57 47 38 66
gwadi 'child'
43
budubadu tetu 'many yams' gum 'bit, small piece' yegila 'name' (drawing)
39 67 58 25
tasisia matala peni 'one has the point of a pencil' kabuluyuvela 'village sector' keda 'road, track, way' vaya 'door' beku 'stone (axe) blade'
A drawing was used, rather than verbal elicitation.
22 48 11 12 7
28
Classificatory Particles in Kilivila
Table 3.1.2. (continued) No. in test
CP to be elicited
Eliciting noun/noun phrase
kala 'day'
yam 'day'
kapwa 'bundles (wrapped up), parcel'
luba 'bundle'
kasa 'row, line'
kasa 'row'
37
kavi 'tool'
kavi 'tool'
49
ke 'wooden things'
kai 'wood, tree'
kila 'cluster, hands of bananas'
usi amakala beya (gesture) 'bananas like this' (arms crossed before the breast)
31 32
3
15
kova Tire'
kova 'fire, fireplace'
68
kubila 'large land plot'
kwabila 'ground, land'
28
kudul 'band of fibers' (at the waistband of a skirt)
doba 'grass skirt'
kudu2 'tooth' (noted as kuduu)*
kudu 'tooth'
59
kumla 'earth oven'
kumkumla 'earth oven'
50
kwe 'thing, anything unknown' (unmarked form for inanimates)
dakuna 'stone'
kweya '(severed) limb'
imitabogu 'my finger'
35
4 8 5
kwoila 'clay pot'
kwena '(Amphlett) pot'
10
kwoya 'mountain, hill'
koya 'mountain, hill'
51
liku 'compartments of a foodhouse'
liku 'foodhouse'
lila 'bough, branch'
keyala deli kelima ketolu 'spear with nine off-branching heads'
13
lilo 'walk, journey'
lola 'journey, trip'
52
24
luva 'wooden dishes'
kaboma 'plate, bowl'
meila 'part of a song'
wosi 'song'
27
mmwa 'conical bundle (of taro)'
budubadu uli 'much taro'
34
nal 'persons of female sex'
vivila 'girl, woman'
na2 'animals' (noted as *bunu)
bunukwa 'pig'
nina 'parts of a song'
wosi kwetala biga 'one word of a song'
26
nunu 'corners of a garden'
nunula bagula 'corner of a garden'
23
nutu 'kneaded things, dot, drop'
notu 'kneaded things, dot, drop'
69 60
9
2 46
oyla 'string, fish on strings'
wela 'string of fish'
pila 'part, piece'
akatui nuya 'I cut a coconut'
18
pona, pwanina 'punctured, hole'
pwanana 'hole'
53
po'ula 'plantation, grove'
po'ula 'plantation, grove'
61
With "noted as" I refer to the fact that I noted down kudo2 as kuduu to differentiate it from kudo1 (which was noted as kudu) in my data. The same holds for the notation of no2 (as bunu). The gloss "(Amphlett-)pot" for kwena refers to the fact that these pots come from the Amphlett Islands.
The System of Classificatory Particles in Kilivila
29
Table 3.1.2. (continued) No. in test
CP to be elicited
Eliciting noun/noun phrase
sa 'nut bunch'
budubadu buva 'many betel nuts' tasisia pikekita tobaki 'one has a bit of tobacco' sipu 'sheaf
si 'small bit' sipu 'sheaf sisi 'bough, cut off, part of a tree' siva 'time' siwa 'sea portions'
16 42 62
sisila 'branch'
to 'basket' telto 'persons of male sex'
peta 'basket' tau, tauwau (drawing) T 'man, men'
6 30 29 63 1
utu 'scrap, parts (cut off)'
babwabutou pikekita ton igau, tasisia ... 'I will cut off a bit of sugar cane, then one has ...'
21
uva 'span'
kuvi amakala (gesture) '(long) kuvi yams tuber like this' (gesture: extending the arms)
41
vili 'untwisted'
tobaki amakala (gesture) 'tobacco like this' (gesture: straightening a stick of tobacco) valu 'village'
vilo 'place, village' wela 'batch/string of fish' ya 'flexible, thin things'
tuta 'time' bwalita 'sea'
basuya yena 'I will string fish' yoyu 'palm branch'
19 14 33
yaml 'day' yam2/yuma 'hand'
yam 'day'
5 54
yama 'her/his hand'
70
yulai'bundles of four things'
tasisia kwevasi buva 'one has four betel nuts' boda (kena yena kena avaka) '(a) group (of fish or so)' tetu 'yams'
yuva 'shoal' 0 'a basketful of yams' (zero classifier)
36 64 40
I approached potential consultants to find out whether they would like to cooperate in the study, using the following statement: (1) magigu kupaisewa deli yegu magi-gu ku-paisewa wish-my 2Ps.-work 'I want you to work with me.'
deli with
yegu I
I then showed them a drawing of two men and elicited an adjective, a numeral, and a demonstrative pronoun in combination with the word for either 'man'
30
Classificatory Particles in Kilivila
or 'men'. I received responses like the following: (2) tomanabweta tau (tauwau) to-manabweta tau male-beautiful man 'beautiful man (men)'
(tauwau) (men)
(3) mtona (mtosina) tau (tauwau) m-to-na (m-to-si-na) this-male-this (this-male-Plural-this) 'this (these) man (men)' (4) tetala tau (teyu tauwau) te-tala tau male-one man 'one man (two men)'
(te-yu (male-two
tau man
(tauwau) (men)
tauwau) men)
Having elicited the CP te/to in this way, I explained that I would like to continue along the same lines, presenting to them different nouns that they would then use with an adjective, a numeral, and a demonstrative pronoun. I did this in the following way: (5) Alivala makala tomanabweta tau, tetala tau, mtona tau. Igau alivala dakuna (. . .), igau yokwa bukulivala . . . a-livala makala to-manabweta tau te-tale tau 1Ps.-say like male-beautiful man male-one man m-to-na this-male-this
tau man
igau then
a-livala IPs.-say
dakuna (. . .) stone (. . .)
igau then
yokwa buku-livala . . . . you 2Ps.Future-say 'I say, beautiful man, one man, this man; then I say, stone (etc.); then you will say . . .'. At this point, I wanted to hear the adjective, numeral, and demonstrative with the CP that agreed with the class of the noun presented (in this example, dakuna kwemanabweta 'beautiful stone', etc.). If the consultant did not react, I said the following: (6) Gala tomanabweta dakuna kena . . . gala to-manabweta dakuna not male-beautiful stone 'It is not male-beautiful stone, but . . .'.
kena . . . . but
I then would pause. If the consultant reacted, the CP he or she produced in connection with the noun was noted down for all three word classes on a separate list; if the consultant did not react, I noted that, too.
The System of Classificatory Particles in Kilivila
31
This method was tested in some informal pretests where it became evident that the best way of eliciting data was the following: 1. To ask for one word class after the other and not attempt to obtain a combination of adjective, numeral, and demonstrative pronoun together with the noun within just one utterance (although such utterances are documented in naturally produced phrases such as mtona doketa tetala toveaka 'this one fat medical doctor'). 2. To ask for the adjective first, then for the numeral, and finally for the demonstrative pronoun. The second point was quite important, especially in working with children. This way of gathering data, using a questionnaire and lists to note the consultants' responses (some of the tests were also tape recorded to document the data collection), as well as the standardized introduction, marked the situation as being quite formal. Nevertheless, my consultants did not seem to be bothered by this; on the contrary, they were highly cooperative and gave no sign of losing interest. Even the children, to whom the test seemed to be a new kind of game, cooperated enthusiastically and attentively. This is quite surprising, considering that it took some consultants up to forty minutes to complete the questionnaire. All sixty tests were done in one session with each consultant. These data constitute the first CP sample for my analyses. The second CP sample consists of all the CPs I found in my overall transcribed corpus of Kilivila language production. I noted who produced the CP, the word class, and the context (see appendix A and appendix B); I also noted the complete noun phrases in which the CPs were produced. The data elicited in the test were ordered and processed for analysis using an IBM PC and ordering programs supplied by the Max-Planck-Institute for Psycholinguistics, Nijmegen, The Netherlands.1 After the analyses were computed (1987-1989), I returned to the field in 1989 to confirm the main results. This restudy, including its aims, methods, and results, is presented in chapter 4. Before presenting the data, I want to make a final, rather general, remark: It should be clear from the methods used to gather the data that the emphasis of my research was on actual language production with respect to the system of CPs in Kilivila. The data and the results presented in the following sections show the system of CPs used in actual speech production by a relatively high percentage of the inhabitants of Tauwema village on Kaile'una Island. 3.2 THE DATA In section 3.2.1, the data are first ordered according to the age group of the consultants, and first interpretations are given. Section 3.2.2 then presents 1 I want to thank Wolfgang Klein, Pim Levelt, Peter Wittenburg, Gerd Klaas, and especially Kees van der Veer for their friendly and generous support. I tried a scalogram analysis (Guttmann) on my data, but recording the data did not produce a typical Guttmann pattern.
Classificatory Particles in Kilivila
32
the CP types and their tokens as they were produced by all consultants during the elicitation test and in the overall corpus of Kilivila speech. This twofold presentation should help readers maintain their orientation in spite of the large amount of data given and it should make it easier for the critical reader to evaluate the results in section 3.3. 3.2.1 Interpretation of the Data by Age Group This section presents the CP data according to the age group of the consultants. First, the data that were elicited with the questionnaire are considered; then, the CP data documented in the overall corpus are presented. Using the questionnaire, I elicited 87 CP types and 10,583 CP tokens. In Table 3.2.0, these tokens are classified according to the consultants' gender and the three word classes.
3.2.1.1 Age Group I Table 3.2.1 gives the number of responses (CP tokens) to the eliciting stimuli presented for demonstrative pronouns, numerals, and adjectives and the number of CP types produced by each consultant in age group I (age range: 4-7 years). The number of responses elicited from the consultants ranged from 4 (consultants 2, 3, and 9) to 164 (consultant 10). Table 3.2.1 confirms that children start to acquire and to produce CPs at approximately the age of 4 (see consultants 2, 3, 9, and 7). The frequency of CPs produced rises rather dramatically at about 6 years of age; however, consultant 1 seems to be an exception to this observation; the only explanation I can offer for her very few responses to the stimuli presented is that Igiobibila is a rather shy little girl indeed. If we compare consultant 5, Kwelubituma, who is almost 6 years old, with consultant 8, Yabilosi, who just passed his fifth birthday, we can conclude that 5 is the critical age with respect to the developing CP acquisition process. Table 3.2.1 also shows that in this age group CPs are most often produced with adjectives, then with numerals, and then with demonstrative pronouns. The frequencies can be interpreted as indices for the order of acquisition of the respective word classes; thus, I tentatively conclude that adjectives are acquired before numerals, and numerals are acquired before demonstrative
Table 3.2.0. CP tokens elicited from 60 consultants using the questionnaire CP/demonstrative
CP/numeral
CP/adjective
Total
Men Women
1,775 1,666
1,807 1,734
1,818 1,783
5,400 5,183
Total
3,441
3,541
3,601
10,583
The System of Classificatory Particles in Kilivila
33
Table 3.2.1. Responses in age group Consultant no. (age) Girls
Demonstrative pronoun
l(6y)
4
6(6y)
2(4y) 3(4y) 4(7y) 5(5y)
Total
Adjective
2 26 18 20
6 1 1 42 42 40
6 1 1 45 50 58
72
132
161
5
6 4 1 52 54
5 5 2 59
2
Boys 7(4y) 8(5y) 9(4y) 10 (6 y) 11 (6y) 12 (7 y)
Numeral
4 1 53 25 53
54
58 54
Total
141
171
183
Total, age group I
213
303
344
No. Of types 8 2 2 18 12 16
6
5
2
29 19 20
pronouns. The consultants of this age group produced only 6.2% of all demonstratives, but 8.6% of all numerals and 9.6% of all adjectives elicited from the 60 consultants in the study. Consultants 4, 5, 6, and 11 produced far fewer demonstrative pronouns than numerals and adjectives, suggesting that demonstratives are more difficult to acquire and to produce than numerals and adjectives. However, these four consultants' decrease in CP production in connection with demonstratives is so dramatic that I cannot completely exclude the interpretation that it may reflect signs of diminishing attention, although this conflicts with the impression I had when I was working with these children. Table 3.2.1 also documents a difference in the CP production of boys and girls. Adding up the production frequencies of all three word classes for girls (72+132+161 = 365) versus boys (141 + 171 + 183 = 495), we find that girls in age group I produced 44.1% of all their responses in connection with adjectives, 36.2% in connection with numerals, and 19.7% in connection with demonstratives, whereas boys in age group I produced 37% of all their responses in connection with adjectives, 34.5% in connection with numerals, and 28.5% in connection with demonstratives. On the basis of these findings, I conclude that the order of acquisition and production of the three word classes is even more marked in girls than in boys. The children of age group I also showed differences with respect to the number of CP types they produced: For girls, the number of types produced ranged from 2 (consultants 2 and 3) to 18 (consultant 4). For the six girls of
34
Classificatory Particles in Kilivila
group I, an average of 7.7 CP types were produced. For boys, the number of types produced ranged from 2 (consultant 9) to 29 (consultant 10). For the six boys of group I, an average of 13.5 CP types were produced. Thus, the boys of this age group not only produced more responses, but they also produced a broader variety of CP types than the girls. This difference may be a result of the Trobrianders' education system, which offers boys a much broader range of possibilities for experiencing their surrounding environment than it offers to girls (see B. Senft 1985). However, the fact that the girls are responding to an adult, foreign, white, male researcher should also be kept in mind. Before proceeding, I want to clarify my use of the term "CP type". At this stage of the data presentation, I assign provisionally to all the CPs listed in the following tables the status "CP type". The subsequent analyses will show which of these CPs are proper CP types and which are merely variants. Table 3.2.2 lists the 37 CP types that were produced by the consultants in age group I, along with a basic gloss (for the meanings of the individual CPs, see Tables 3.1.2 in section 3.1 and 3.3.1 in section 3.3.1). As shown in Table 3.2.2, the CPs to/te and nal were produced by all consultants; obviously, they are the first CPs that are acquired. The CPs ke and kwe were produced by 4 girls and 5 boys; thus, it can be assumed that these two CPs are acquired next. The CP kumla was produced by 4 girls and 3 boys; pwanina/pona was produced by 3 girls and 4 boys; kova, na2, ta, and ya were produced by 3 girls and 3 boys; ka'i was produced by 3 girls and 2 boys; and sisi was produced by 2 girls and 3 boys. Thus, I infer that these CPs are acquired after to/te, nal, ke, and kwe. Only girls produced the CPs kwoya, kaduyo, lilo, va, and yuma. On the other hand, 4 boys, but only 1 girl produced the CP gudi; moreover, only boys produced the CPs bubwa, buda, bukwa, duli, kai, kapwa, kweya, kwoila, luba, luva, notu, pila, and sa. Again, this seems to reflect the broader range of experiences for boys compared with their female peers. With all other CPs produced by the consultants in age group I, there were no striking asymmetries. However, the list documents the rather dramatic rise in CP production at about 6 years of age. Table 3.2.3 shows the CP types ordered according to the number of tokens produced by the 12 consultants. It confirms the interpretations given for the data in Table 3.2.2; with the exception of kada, all the CP types that I propose are acquired in the first three stages of the CP acquisition process were produced most frequently (i.e., had the largest number of tokens). Consultants 6 and 12 produced the CP kada as a response to the stimuli presented to elicit the CPs kada, duya, and kaduyo; this explains the relatively high number of tokens (17) for this CP type. Thus, to understand the rather high number of tokens for some CP types, we must consider which CPs were expected compared with which CPs were actually produced by the consultants. Table 3.2.4 presents the result of this comparison. A full discussion of the data presented in Table 3.2.4 will be given in section
The System of Classificatory Particles in Kilivila
35
Table 3.2.2. CP types/tokens produced by age group No. of girls/boys producing CP type
Tokens produced by girls/boys
bubwa 'part cut off' buda 'group' bukwa 'fruit cluster'
0/1 0/2 0/1
0/1 0/8 0/3
bwa 'tree, wood' duli 'cluster' gudi 'child'
1/1 0/1 1/4
9/1 0/3 1/6
iga 'name' kada 'road'
2/2 1/2
4/6 5/2
kaduyo 'entrance'
1/0
kai 'stone blade' ka'i 'tooth' kapwa 'parcel'
0/1 3/2
1/0 0/3
CP type
kasa 'row' ke 'wood' kova 'fire' kumla 'earth oven' kwe 'thing'
0/2 1/3 4/5 3/3 4/3
8/5 0/5 3/8 85/119 9/8 10/9
134/171
kweya 'limb'
4/5 0/1
kwoila 'clay pot' kwoya 'mountain'
0/2 1/0
lilo 'walk' luba 'bundle'
1/0 0/1
luva 'wooden dishes'
0/1
0/1
nal 'female person' na2 'animals' notu 'kneaded, dot'
6/6 3/3
15/18 22/18
0/2 0/3 3/4
0/3 0/11 7/10 0/4
pila 'part, piece' pwanina, pona 'hole' sa 'nut bunch' si 'small bit' sisi 'bough' ta 'basket' tolte 'male person' utu 'scrap' va 'door, window' va 'flexible, thin' yuma 'hand'
0/1 1/2 2/3 3/3 6/6 2/2 1/0 3/3 1/0
0/3 0/5 1/0 1/0 0/1
1/4 4/8 10/9 19/19 4/4 1/0 10/12 1/0
36
Classificatory Particles in Kilivila
Table 3.2.3. Number of tokens produced for each CP type—age group I Type
Tokens
kwe 'thing' na2 'animals' nal 'female' kumla 'earth oven' kada 'road' pwaninalpona 'hole' sisi 'bough' pila 'part' iga 'name' utu 'scrap' kapwa 'parcel' si 'bit' duli 'cluster' kweya 'limb' bubwa 'cut off kwoya 'mountain' luba 'bundle' sa 'nut bunch' yuma 'hand'
305 39 34 19 17 17 12 11 10 8 5 5 3 3 1 1 1 1
Type ke 'wood' tolte 'male' ya 'flexible' ta 'basket' kova 'fire' ka'i 'tooth' kasa 'row' bwa 'tree' buda 'group' gudi 'child' kwoila 'pot' bukwa 'cluster' kai 'blade' notu 'kneaded' kaduyo 'entrance' lilo 'walk' luva 'dishes' va 'door'
Tokens 204 38 22 19 17 13 11 10
1
Types total: 37; tokens total: 860
Table 3.2.4. CPs produced versus CPs expected—age group I Produced
Expected
bubwa 'cut off buda 'group' bukwa 'cluster'
utu 'scrap' buda 'group', yuva 'shoal' bukwa 'cluster'
bwa 'tree'
bwa 'tree', bubwa 'cut off, kabisi 'compartment/foodhouse', ke 'wood', utu 'scrap' duli 'cluster' gudi 'child' iga 'name' kada 'road', duya 'door', kaduyo 'entrance' kai 'stone blade' kudu2 'tooth' kapwa 'parcel' kasa 'row', iga 'name'
duli 'cluster' gudi 'child' iga 'name' kada 'road' kai 'stone blade' ka'i 'tooth' kapwa 'parcel' kasa 'row'
8 7 5 3 3 3 1 1 1 1
The System of Classificatory Particles in Kilivila
37
Table 3.2.4. (continued) Produced
Expected
ke 'wood'
ke 'wood', bubu 'block cut off, bubwa 'cut off', bukwa 'cluster' bwa 'tree', duya 'door', giwi 'cut' kabisi 'compartment/foodhouse', kabulo 'point', kabulu 'village sector', kasa 'row', kila 'hands of bananas', lila 'bough', lilo 'walk', luva 'dish', mmwa 'bundle', oyla 'string', pila 'part', sa 'nut bunch', si 'bit', sipu 'sheaf, utu 'scrap', uva 'span', vili 'untwisted', ya 'thin', yulai 'bundle of four things', yuva 'shoal'
kova 'fire'
kova 'fire', yuma 'hand'
kumla 'earth oven'
kumla 'earth oven'
kwe 'thing'
kwe 'thing', bogi 'night', buda 'group', deli 'group', duli 'cluster', gula 'heap', gum 'bit', iga 'name', kabisi 'compartment/ foodhouse', kala 'day', kapwa 'parcel', kila 'hands of bananas', kubila 'land plot', kudul 'fibers', kwoila 'clay pot', kwoya 'mountain', liku 'compartments/foodhouse', lilo 'walk', meila 'song part', nina 'song part', notu 'kneaded', nunu 'corner/ garden', po'ula 'grove', pwanina 'hole', sa 'nut bunch', sipu 'sheaf, siva 'time', siwa 'sea portion', uva 'span', vilo 'place', yam 'day', yulai 'bundle of four things', yuma 'hand', yuva 'shoal', 0 'basketful of yams'
kweya 'limb'
kweya 'limb'
kwoila 'clay pot'
kwoila 'clay pot'
kwoya 'mountain'
kwoya 'mountain'
lilo 'walk'
lilo 'walk'
luba 'bundle'
kapwa 'parcel'
luva 'dish'
luva 'wooden dishes'
nal 'female'
nal 'female persons', bubu 'cut across', deli 'group', kweya 'limb', uva 'span'
na2 'animals'
na2 'animals', gum 'bit', wela 'batch of fish'
notu 'kneaded'
notu 'kneaded, dot'
pila 'part'
pila 'part', kabisi 'compartment/foodhouse', kubila 'land plot', meila 'song/part'
pwaninalpona 'hole'
pwanina 'hole'
sa 'nut bunch'
sa 'nut bunch'
si 'small bit'
si 'small bit', ya 'flexible, thin'
sisi 'bough'
sisi 'bough', ya 'flexible, thin'
ta 'basket'
to 'basket', giwi 'cut'
tolte 'male'
tolte 'male persons', deli 'group', yuma 'hand'
utu 'scrap'
utu 'scrap', si 'small bit', vili 'untwisted'
va 'door' ya 'flexible'
kaduyo 'entrance' ya 'flexible, thin', bubu 'block cut off, gili 'row', pila 'part', sisi 'bough'
yuma 'hand'
yuma 'hand'
38
Classificatory Particles in Kilivila
3.3. However, at this point, the following can be noted: 1. kasa 'row/line' was produced as a response to the stimulus for the expected CP iga 'name'. This is not acceptable in adult Kilivila; the CP kasa does not agree with the noun yegila 'name'. 2. Kova 'fire' was produced as a response to the stimulus for the expected CP yuma 'hand'. This is not acceptable; kova does not agree with the noun yamala 'her/his hand'. 3. nal 'female persons' was produced as a response to the stimuli for the expected CPs bubu 'block cut off, kweya 'limb', and uva 'span'. This is not acceptable; nal does not agree with the nouns bobu 'log, round timber', kuvi 'type of long yams', and imitabogu 'my finger' (here the reference is unequivocal: the consultant producing this CP and the person eliciting this CP were both male). 4. Na2 'animals' was produced as a response to the stimulus for the expected CP gum 'bit'. This is not acceptable; na2 does not agree with the noun gum 'bit, small piece'; however, gum is also the name for a snail-like ornament on a canoe board. If consultant 5 knew this rather special meaning, then na2 would agree with the noun, presuming that the consultant wanted to refer to the snail in the ornament. However, this explanation is rather unlikely. 5. Ta 'basket' was produced as a response to the stimulus for the expected CP giwi 'cut'. The consultant may have confused the stimulus giwi 'cut' with givi. Givi translates as 'serving of fish', which is often presented on a small, woven, fiber mat. The consultant may have thought of this mat when she produced this (then acceptable) response. If we correlate these observations with the individual consultants, we note that one dubious case of CP production occurred with consultants 4 and 11; with consultant 5, there were 7 dubious cases of CP production. Therefore, Table 3.2.4 demonstrates that speakers do make mistakes in the production of CPs and that they show linguistic insecurity with respect to the appropriate usage of some CPs. This becomes more evident if we look at the cases where individual consultants produced more than one CP type as a response to a given stimulus. Table 3.2.5 presents the responses documenting such multiple classification. It will be discussed in greater detail in section 3.3. I will now present the CPs documented in the overall corpus of Kilivila speech data that were produced by children of the same age as the consultants in age group I. Table 3.2.6 lists the consultants (see appendix A) and the number of CP types and tokens for demonstrative pronouns, numerals, and adjectives. Table 3.2.6 presents a somewhat different picture from that in Table 3.2.1. However, Table 3.2.6 does not provide any counterevidence for the interpretation of the data in Table 3.2.1 with respect to the order of acquisition of CPs with the three word classes. As documented in appendix B, the
The System of Classificatory Particles in Kilivila
39
Table 3.2.5. Responses documenting multiple classification in age group I Consultant no.
Cases of multiple classification
Percent
0 0 0 5 3 5
0 0 0 28 25 31
Girls 1
2 3 4 5 6
(Average: 14) Boys
0 0 0 6 4 1
7 8
9 10 11 12
0 0 0 21 21 5
(Average: 8) No. of cases of multiple classification in relation to all CP types produced by the consultant.
Table 3.2.6. Consultants and CP types and tokens—overall corpus of Kilivila speech data Consultant no. (age) Boys 11/8 (5y) 12/10 (6 y) 13 (7y) Total
Adjective
No. Of types
15 7
0 15 8
10 7
23
23
Demonstrative pronoun
Numeral
0 64 68 132
1
1
relatively high frequency of CP tokens produced with demonstratives depends on the text category in which the CPs are produced: 125 of the 132 CP tokens with demonstrative pronouns were produced in the text category "fairly tale". This offers the interesting possibility that fairy tales are told not only to entertain children and to transmit cultural ideas and values, but also as a means to foster the language acquisition process (see also Erbaugh 1986:412). The number of CP types that the three boys in the overall corpus produced ranges from 1 to 10. Table 3.2.7 presents the types and the tokens produced by consultants 11/8, 12/10, and 13. It supports, at least in part, the interpretation of Tables 3.2.2 and 3.2.3. The CPs nal, to/te, na2, ke, kwe, and
40
Classificatory Particles in Kilivila
Table 3.2.7. CP types and tokens produced by the three consultants in the overall corpus of Kilivila speech data Type
Demonstrative pronoun
Numeral
Adjective
Total
bukwa 'cluster'
4
0
dumia 'swamp' kada 'road' kai 'blade'
0
0
5
0
0
1
0
ke 'wood' kova 'fire' kwe 'thing' nal 'female' na2 'animals' pila 'part' pwanina 'hole' tolte 'male' ya 'thin'
17 2
4 0 4
3 0 6 4
19 43
7 2
2
28
0
0
0
2 1
27 10
3 1
0
35 11
132
23
23
178
Total
9 38 19 0 1
1
2 1
6 1
0
5 1
5
24 2
ya are acquired in the first stages of the CP acquisition process. The CP dumia 'swamplike' was produced by consultant 12/10 as part of the adjective dumiaveaka, but without a noun to complete the noun phrase; the noun phrase using this adjective translates as 'big swamplike ground'. There was no peculiar usage of any of these CPs with respect to the noun they referred to and agreed with. However, the following facts should be noted: 1. Consultant 12/10 produced the CP bukwa 'fruit cluster' with an adjective and a demonstrative pronoun to refer to the noun bwaibwai 'green coconut', which he introduced three and seven sentences before the respective references. 2. Consultant 13 not only played with the contextually adequate usage of the CPs to/te 'male persons' and nal, but also showed two cases of self-correction; he started the following phrase twice with a CP that did not agree with the noun: makadana . . . mmna . . . manakwa koya 'this(*)road-this . . . hm . . . this-thing-this mountain'. In summary, the following observations can be made about the acquisition and production of CPs in consultants ranging in age from 4 to 7 years: 1. Children start to acquire and produce CPs at the age of 4; 5 is the critical age with respect to CP acquisition. At about 6 years, there is an explosion in children's production of CPs.
The System of Classificatory Particles in Kilivila
41
2. If the use of specific CPs in certain text categories is ignored, the general tendency for this age group seems to be that CPs are most often produced with adjectives, then with numerals, and then with demonstrative pronouns. These frequencies are interpreted as indices for the order of the acquisition and production of the respective word classes. This order is more marked in girls than it is in boys. 3. Boys produced more CP types and tokens than girls. 4. The first CPs that are acquired are to/te and nal; ke and kwe come second; and kumla, ka'i, kova, na2, ta, ya, pwanina/pona, sisi, and (at least for boys) gudi come third. 5. Some of the CPs produced in this age group did not agree with the nouns or noun phrases they referred to. In addition to these cases of dubious or not acceptable CP usage, the production of quite a few of the CPs showed linguistic insecurity in some of the consultants (i.e., they used a CP to refer to a given nominal donatatum inappropriately). 3.2.1.2 Age Group II Table 3.2.8 gives the number of responses (CP tokens) to the eliciting stimuli presented for demonstrative pronouns, numerals, and adjectives and the number of CP types produced by each consultant in age group II (8-14 years). The number of responses elicited from the consultants ranged from 126 (consultant 14) to 210 (consultants 13, 18, and 20). However, it should be Table 3.2.8. Responses in age group II Consultant no. (age)
Demonstrative pronoun
Numeral
Adjective
No. of types
Girls 70 31 63 63 69 70
70 38 63 63 69 70
70 57 64 63 69 70
366
373
393
66 70 68 69 64 65
67 70 68 69 65 65
66 70 68 69 65 65
Total
402
404
403
Total, age group II
768
777
796
13 (11y)
14 (8y)
15 (11y) 16 (11y)
17 (12y) 18 (14 y)
Total Boys 19(13y) 20 (9y) 21 (12 y) 22 (14 y) 23 (10 y) 24 (8y)
22 16 33 21 32 24
34 29 35 29 21 31
42
Classificatory Particles in Kilivila
noted that consultant 14, Olopola, just passed her eighth birthday. If we compare the number of her responses with those produced by consultant 24, Luluwasikweguyau, who is almost 9 years old, and with the number produced by consultants of age group I (Table 3.2.1), the following interpretation seems plausible: The number of responses Olopola produced may indicate that 9 is another critical age with respect to the CP acquisition process. The production rate of CPs obviously increases at about 9 years of age. Thus, we can modify our first comment on Table 3.2.8 in the following way: Leaving aside consultant 14, the number of responses elicited in the other 11 consultants of age group II ranges from 189 (consultant 16) to 210 (consultants 13, 18, and 20). There are no apparent differences with respect to the order of production of CPs with the three word classes, again excluding consultant 14. Thus, I infer that 9 is indeed not only a critical age with respect to the CP acquisition process from the point of view of CP production frequency, but also from the point of view of the balanced use of CPs with adjectives, numerals, and demonstrative pronouns. The consultants of this age group produced 22.3% of all demonstatives, 21.4% of all numerals, and 22.1% of all adjectives elicited from the 60 consultants in the study. Boys in age group II produced only a few more CPs than girls, and this difference becomes negligible if consultant 14 is excluded from the interpretation of Table 3.2.8 and if the average frequency of CP tokens produced is calculated for the remaining five girls (demonstratives = 67 CP tokens; numerals = 67 CP tokens, adjectives = 67.2 CP tokens) and for the boys (demonstratives = 67 CP tokens, numerals = 67.3 CP tokens, adjectives = 67.2 CP tokens). With respect to the number of CP types produced by boys and girls the data show the following: The number of CP types produced by the girls of age group II (again, excluding consultant 14) ranged from 21 (consultant 16) to 33 (consultant 15). For the six girls of age group II, an average of 24.7 CP types were produced; if we exclude Olopola once more, the average is 26.4 CP types produced. With the boys, the number of CP types produced ranged from 21 (consultant 23) to 35 (consultant 21). For the six boys of this age group, there were an average of 29.8 CP types produced. Thus, the boys of this age group produced only a few more CP tokens than the girls as responses to the stimuli presented; however, it should be noted that the boys still produced a broader variety of CP types than the girls did. Again, this difference may be due to differences in the Trobrianders' education system, which offers boys a much broader range of possibilities to experience their surrounding environment. Table 3.2.9 presents the 57 CP types that were produced by the consultants of age group II. The CPs iga, kada/keda, ke, kova, kumla, kwe, nal, na2, sisi, ta, and to/te were produced by all consultants. The CP utu was produced by 6 girls and 5 boys; ya was produced by 5 girls and 6 boys; and gudi was produced by 5 girls and 5 boys. Thus, I infer that the process of acquiring
The System of Classificatory Particles in Kilivila
43
Table 3.2.9. CP types/tokens produced by age group II No. of girls/boys producing CP type
Tokens produced by girls/boys 0/2 3/4
buda 'team'
0/1 1/2 5/4
12/8
bwa 'tree'
3/3
14/12
deli 'group'
1/3
duli 'cluster'
1/1 1/4
3/12
giwi 'cut'
2/5
4/13
gudi 'child'
5/5
15/15
gula 'heap'
5/3
gum 'bit'
1/1 2/4
16/18
CP type bogi 'night' bubu 'cut across'
3/6
iga 'name'
6/6
(*) iki '???'
1/0
1/0
kabisi 'compartment'
2/2
4/6
kabulu 'sector'
0/1 6/6
32/49
kadalkeda 'road'
0/2
kaduyo 'entrance'
1/0
1/0
kai 'blade'
4/6 3/6
10/17
3/3 3/5
9/9
6/6 1/0
232/267
6/6 1/3 1/0 6/6 6/6
18/18
18/18 438/366
3/5 5/4
15/11
ka'i 'tooth' kapwa 'parcel' kasa 'row' ke 'wood' kila 'cluster' kova 'fire' kubila 'land plot' kudu 'tooth' kumla 'earth oven' kwe 'thing' kweya 'limb' kwoila 'clay pot' kwoya 'mountain' liku 'compartment' luba 'bundle' luva 'wooden dish' megwa 'magic' mmwa 'bundle' nal 'female' na2 'animals'
1/1 2/3 1/2 1/0 0/1 0/1 6/6 6/6
9/21 9/23
2/0 3/9 3/0
9/11
1/3 2/9 3/6 3/0 0/1 0/2 21/18 28/39
In tables, (*) indicates that, according to my other consultants, there is no such CP in Kilivila.
44
Classificatory Particles in Kilivila
Table 3.2.9. (continued) CP type notu 'kneaded/dot' nunu 'corner' pila 'part' po'ula 'grove' pwanina 'hole' sa 'nut bunch'
No. of girls/boys producing CP type
Tokens produced by girls/boys
1/5 1/0 3/4
3/9 1/0
0/1 4/6 0/2
17/22
0/3 8/16
0/3 5/0
sipu 'sheaf
2/0 1/4
sisi 'bough'
6/6
30/32
suya 'batch/fish'
2/0 6/6
16/18
tolte 'male'
0/3 6/6
24/22
utu 'scrap'
6/5
21/11
va 'door'
2/0 0/1
7/0 0/1 3/0
si 'small bit'
ta 'basket' tarn 'sprouting'
vili 'untwisted'
1/10
6/0 0/10
wela 'batch/fish'
1/0
ya 'thin'
43/45
yam 'day'
5/6 0/2
yuma 'hand'
0/1
0/2
0/4
these CPs has been almost completed in this age group. Six boys but only 4 girls produced the CPs kai and pwanina; 6 boys but only 3 girls produced the CP ka'i. Thus, I infer that the boys have acquired these CPs, while the girls are still in the process of acquiring them. Five girls and 4 boys produced the CPs kwoila and buda; 5 boys and 3 girls produced the CPs kasa and kweya. I infer that the process of acquiring these CPs has been almost completed. Five boys but only 2 girls produced the CP giwi, and 5 boys but only 1 girl produced the CP notu. Thus, I infer that boys acquire these two CPs earlier than girls. This seems to hold for the CPs duli, sipu, and gum, too: 4 boys but only 1 girl produced duli and sipu, and 4 boys but only 2 girls produced gum. The CP pila was produced by 4 boys and 3 girls; bwa and kapwa were produced by 3 boys and 3 girls. Boys seem to acquire the CPs kubila, liku, and tam earlier than girls: 3 boys produced these CPs, whereas only 1 girl produced the CP kubila, two girls produced the CP liku, and no girls produced the CP tam. Consultant 15 produced the CP (*)iki; however, my other consultants told me that there is no such CP in Kilivila. Table 3.2.10 presents the CP types ordered according to the number of
The System of Classificatory Particles in Kilivila
45
Table 3.2.10. Number of tokens produced for each CP type—age group II Type kwe 'thing' ya 'thin' na2 'animals' to/te 'male' pila 'part' kova 'fire' ta 'basket' utu 'scrap' ka'i 'tooth' bwa 'tree' pwanina 'hole' kweya 'limb' giwi 'cut' kubila 'land plot' liku 'compartment' kabisi 'compartment gum 'bit' gula 'heap' va 'door' si 'small bit' kwoya 'mountain' kudu 'tooth' po'ula 'grove' wela 'batch/fish' kabulu 'sector' mmwa 'bundle' (*) iki '???' megwa 'magic' vili 'untiwsted'
Tokens 804 88 67 46 39 36 34 32 30 26 24 20 17 12 11 10 9 8 7 5 4 3 3 3 2 2 1 1
Type ke 'wood' kadalkeda 'road' sisi 'bough' nal 'female' kumla 'earth oven' iga 'name' kasa 'row' gudi 'child' km 'blade' kwoila 'clay pot' buda 'group' kapwa 'parcel' duli 'cluster' notu 'kneaded, dot' sipu 'sheaf tam 'sprouting' luba 'bundle bubu 'cut across' suya 'batch/fish' deli 'group' yam 'day' luva 'wooden dish' sa 'nut bunch' bogi 'night' kila 'cluster' yuma 'hand' kaduyo 'entrance' nunu 'corner/garden'
Tokens 499 81 62 39 36 34 32 30 27 26 20 18 15 12 11 10 9 7 6 4 4 3 3 2 2 2 1 1
1
Types total: 56 (57); tokens total: 2,340 (2,341)
tokens realized by all 12 consultants. Table 3.2.11 lists the CPs produced versus those expected in response to the stimuli presented. These data will be discussed in detail in section 3.3. However, the following should be noted here: 1. Buda 'group' was produced as a response to the stimulus presented to elicit the expected CPs bubu 'block cut across/cut off and gum 'bit, small piece'. This is not acceptable; buda does not agree with the nouns bobu 'log, round timber' and gum 'bit, small piece'.
46
Classificatory Particles in Kilivila
Table 3.2.11. CPs produced versus CPs expected—age group II Produced
Expected
bogi 'night' bubu 'cut across" buda 'group, team' bwa 'tree'
bogi 'night' bubu 'block cut across/cut off buda 'group', bubu 'cut across', gum 'bit', yuva 'shoal' bwa 'tree, wood', bubwa 'parts cut off, kabisi 'compartment/ foodhouse', kabulo 'village sector', ke 'wooden things', si 'small bit', utu 'scrap', vili 'untwisted' deli 'group' duli 'cluster, bundle' giwi 'cut' gudi 'child' gula 'heap, group', bubu 'cut across' gum 'bit' iga 'name' nina 'parts/song' kabisi 'compartment/foodhouse'
deli 'group' duli 'cluster' giwi 'cut' gudi 'child' gula 'heap' gum 'bit' iga 'name' (*) iki'???' kabisi 'compartment' kabulu 'sector' kada/keda 'road' kaduyo 'door' kai 'stone blade' ka'i 'tooth' kapwa 'parcel' kasa 'row' ke 'wooden things'
kila 'cluster' kova fire' kubila 'land plot' kudu 'tooth' kumla 'earth oven' kwe 'thing'
kabulu 'village sector' kada 'road, track', duya 'door, entrance', kaduyo 'door, entrance', lilo 'walk', vilo 'place, village' kaduyo 'door, entrance' kai 'stone blade' kudu 'tooth', oyla 'string' kapwa 'parcel, bundle' kasa 'row', gili 'row', giwi 'cut', gum 'bit', yuva 'shoal' ke 'wooden things, tree', bubwa 'cut across', bukwa 'fruit cluster', bwa 'tree, wood', duli 'cluster, bundle', giwi 'cut', kabisi 'compartment/foodhouse', kabulo 'point', kabulu 'village sector', kaduyo 'door, entrance', kavi 'tool', kila 'cluster/bananas', kudu 'tooth', kweya 'limb', kwoila 'clay pot', liku 'compartment/ foodhouse', lila 'bough', luva 'wooden dishes', meila 'song/part', mmwa 'conical bundle', oyla 'string', pila 'part', sa 'nut bunch', si 'small bit', sisi 'bough', utu 'scrap', uva 'span', vili 'untwisted', ya 'thin', yulai 'bundle of four things' kila 'cluster/bananas' kova fire' kubila 'land plot' kudu 'tooth' kumla 'earth oven' kwe 'thing', bogi 'night', bubu 'cut across', buda 'group', bukwa 'cluster', deli 'group', duli 'cluster', duya 'entrance', gili 'row', giwi 'cut', gula 'heap', gum 'bit', kabisi 'compartment/ foodhouse', kabulu 'village sector', kaduyo 'door', kai 'stone blade', kala 'day', kapwa 'parcel', kasa 'row', kavi 'tool', kila 'cluster/bananas', kubila 'land plot', kudu 'band of fibers', kudu
The System of Classificatory Particles in Kilivila
47
Table 3.2.11. (continued) Produced
Expected 'tooth', kweya 'limb', kwoila 'clay pot', kwoya 'mountain', liku 'compartment/foodhouse', lilo 'walk', luva 'wooden dishes', meila 'song/part', mmwa 'conical bundle', nina 'song/part', notu 'kneaded, dot', nunu 'corner/garden', oyla 'string', po'ula 'grove', pwanina 'hole', sa 'nut bunch', si 'small bit', sipu 'sheaf, siva 'time', siwa 'sea portion', uva 'span', vili 'untwisted', vilo 'place', yam 'day', yulai 'bundle of four things', yuma 'hand', yuva 'shoal', 0 'basketful of yams'
kweya 'limb'
kweya 'limb'
kwoila 'clay pot' kwoya 'mountain'
kwoila 'clay pot' kwoya 'mountain'
liku 'compartment' luba 'bundle'
liku 'compartment/foodhouse' kapwa 'parcel, bundle'
luva 'wooden dishes' megwa 'magic' mmwa 'conical bundle'
giwi 'cut' nina 'song/part' mmwa 'conical bundle'
nal 'female' na2 'animals'
nal 'female persons', kai 'stone blade' kweya 'limb' na2 'animals', giwi 'cut', oyla 'string', wela 'batch of fish'
notu 'kneaded, dot' nunu 'corner/garden'
notu 'kneaded, dot' kabisi 'compartment/foodhouse'
pila 'part'
pila 'part', kavi 'tool', meila 'song/part', nina 'song/part', ya 'thin'
po'ula 'grove'
po'ula 'grove'
pwanina 'hole' sa 'nut bunch'
pwanina 'hole' sa 'nut bunch' sipu 'sheaf ya 'thin' sisi 'bough', kai 'stone blade', ya 'thin'
sipu 'sheaf si 'bit' sisi 'bough' suya 'batch of fish' ta 'basket' tam 'sprouting yams' tolte 'male' utu 'scrap'
va 'door' vili 'untwisted' wela 'batch of fish' ya 'thin'
yam 'day' yuma 'hand'
wela 'batch of fish' to 'basket' gula 'heap', 0 'basketful of yams' tolte 'male', deli 'group' utu 'scrap', gili 'row', gum 'bit', kabisi 'compartment/foodhouse', kabulo 'point', kabulu Village sector', kavi 'tool', vili 'untwisted', si 'small bit' kaduyo 'entrance' vili 'untwisted' wela 'batch of fish' ya 'flexible, thin', bubu 'cut across', bukwa 'fruit cluster', duya 'door', gili 'row', gum 'bit', kaduyo 'entrance', lilo 'walk', pila 'part', sipu 'sheaf yam 'day' yuma 'hand'
48
Classificatory Particles in Kilivila 2. Gula 'heap, group' was produced as a response to the stimulus for the expected bubu 'block cut across/cut off'. This is not acceptable; gula does not agree with the noun bobu 'log, round timber'.
3. Consultant 15 produced iki as a response to the stimulus for the expected nina 'parts of a song'; all my other consultants told me that there is no CP iki in Kilivila. 4. Kada/keda 'road, track' was produced as a response to the stimuli for the expected CPs lilo 'road, journey' and vilo 'place, village'. The stimulus lola 'journey, trip' may also elicit the CP kada/keda, although this understanding of the stimulus is somewhat idiosyncratic. The CP kada/keda, however, does not agree with the eliciting noun valu 'village' for the CP vilo; thus, this response is not acceptable. 5. Ka'i 'tooth' was produced as a response to the stimulus for the expected CP oyla 'string'. This is not acceptable; the CP ka'i does not agree with the stimulus wela presented to elicit the CP oyla. 6. Kasa 'row, line' was produced as a response to the stimuli for the expected CPs giwi 'cut', gum 'bit', and yuva 'shoal'. Though this reaction is rather dubious, the use of kasa as a response to the eliciting words can be explained: if we concede that the consultants may have thought of something cut in rows, of bits or pieces presented in a row, or of a group or shoal of fish swimming in a row, we can accept this response as appropriate. 7. Luva 'wooden dishes' was produced as a response to the stimulus giwi 'cut' for the expected CP giwi 'cut'; this is not acceptable. 8. Nal 'female persons' was realized as a response to the stimuli for the expected CPs kai 'stone blade' and kweya 'limb'. This is not acceptable; the CP nal agrees with neither the noun beku 'stone axe blade' nor the noun phrase imitabogu 'my finger'. 9. Na2 'animals' was produced as a response to the stimulus presented for the expected CP giwi 'cut'. This is not acceptable. However, it may be that the consultant misunderstood the stimulus giwi 'cut' as givi 'a serving of fish'; this misunderstanding would explain the production of the CP na2. 10. Nunu 'corner of a garden' was produced as a response to the stimulus for the expected CP kabisi 'compartment of a foodhouse, of a canoe'. This is not acceptable. 11. Pila 'part, piece' was realized as a response to the stimulus for the CP kavi 'tool'. This is rather dubious, if not unacceptable. 12. Sisi 'bough' was produced as a response to the stimulus for the expected CP kai 'stone blade'. This is not acceptable; sisi does not agree with the noun beku 'stone axe blade'. 13. Utu 'scrap, parts cut off was produced as a response to the stimuli for the expected CPs gili 'row', kavi 'tool', and vili 'untwisted'. Although
The System of Classificatory Particles in Kilivila
49
the use of utu as a response to the phrase given to elicit the CP vili can be explained, the use of utu as a response to the stimuli presented to elicit the CPs gili and kavi is not acceptable; utu does not agree with the nouns vakala 'belt of spondylus shell discs' and kavi 'tool'. If we relate these observations to the individual consultants, we find the following: With consultants 15 and 22, there were 2 dubious cases of CP production. With consultant 14, there were 4 dubious cases of CP production. Consultants 13 and 17 each produced 5 cases of dubious CP production, and consultant 23 produced 12 dubious cases. Thus, Table 3.2.11 demonstrates that some consultants of age group II do make mistakes in the production of the acquired CPs and that they show linguistic insecurity with respect to the appropriate use of some CPs. This linguistic insecurity becomes more evident if we look at the cases where individual consultants produced more than one CP type as a response to the stimulus given. Table 3.2.12 presents the responses documenting such multiple classification; it will be discussed in more detail in section 3.3. Table 3.2.13 lists the number of CP types and tokens for demonstrative pronouns, numerals, adjectives, and interrogative pronouns in the overall corpus of Kilivila speech data that were produced by children and adolescents of the same age as the consultants in age group II in the elicitation test (see also appendix A). The data in this table depend heavily on the text category in which the CPs were produced (see appendix B). Consultant 111/19 produced
Table 3.2.12. Responses documenting multiple classification in age group II Consultant no.
Girls 13 14 15 16 17 18
Cases of multiple classification
Percent
11 6 9 6 7 6
50 38 27 29 22 25
(Average: 32) Boys 19 20 21 22 23 24
10 11 9 6 9 4
29 38 26 21 43 13 (Average: 28)
No. of cases of multiple classification in relation to all CP types produced by the consultant.
50
Classificatory Particles in Kilivila
Table 3.2.13. Consultants and CP types and tokens—overall corpus of Kilivila speech data Consultant no. (age) Boys 111/19 (13 y) 112/20 (9y) 113 114
(9y) (12 y)
Total
Demonstrative pronoun
Numeral
Adjective
Interrogative pronoun
No. Of types
7 41 7
2 2 17 1
4 0 10 1
0 2 0 0
4 4 6 1
56
22
15
2
1
his CP tokens in the process of giving a route description (see G. Senft 1986:132,145-153); consultant 112/20 produced his CP tokens in telling fairy tales; consultant 113 produced his CP tokens during an interview about a ritualized children's fight and in connection with fairy tales. His rather abundant use of demonstrative pronouns is due to the references necessary to secure text coherence in Kilivila fairy tales (this observation supports the interpretation of Table 3.2.6); consultant 114 also produced his CP tokens during the interview about a ritualized children's fight. The number of CP types produced by the four boys in this corpus ranges from 1 (114) to 6 (113). Table 3.2.14 presents the types and the tokens produced by consultants 111/19, 112/20, 113, and 114. This table supports, at least in part, the interpretation of Tables 3.2.9 and 3.2.10 (see also Table 3.2.7). There was no peculiar usage of any of these CPs with respect to the nouns they referred to and agreed with. However, the following facts should be noted: Consultant 113 produced three English numerals (see G. Senft 1987b:342, ftn. 5), two of them with a demonstrative (e.g., mtosina seven dokonikani 'these
Table 3.2.14. CP types and tokens produced by the consultants in the overall corpus of Kilivila speech data Type kabulo 'sector' ke 'wood' kwe 'thing' nal 'female' na2 'animals' pila 'part' to/te 'male' ya 'thin' Total
Demonstrative pronoun 0 7 4 4 4 0 37
Numeral 3 5 3 2 1 1
Adjective 0
1
Interrogative pronoun 0 0
Total 3 13 11 10 6 1
0
0
7 0
4 3 1 0 4 1
0
49 1
56
22
14
2
94
1
0 0
1
The System of Classificatory Particles in Kilivila
51
seven monsters'); however, he used the appropriate CP he to refer with a numeral to the loanword selinboti 'sailing boat' in the phrase ketala selinboti 'one sailing boat' and the adequate CP kwe to refer with a numeral to the English word store in the phrase kwetala store 'one store'. Moreover, this consultant produced a phrase with a double classification of the loanword poseni 'poison', namely, manakwa sitana keta poseni '(thing)-this bit of (wooden material)-one poison'. It should be clear that these observations in connection with the overall corpus of Kilivila speech data are anecdotal because of the rather small database of this sample. In summary, the following observations apply to the consultants in age group II with respect to their acquisition and production of CPs: 1. I tentatively infer from the data that 9 years old is another critical age with respect to the CP acquisition process; the frequency of CPs produced increases rather dramatically at age 9. 2. Age 9 is also a critical age with respect to the CP acquisition from the point of view of a balanced number of CP tokens produced with adjectives, numerals, and demonstrative pronouns, with the exception of the use of specific CPs in certain text categories. 3. Boys produced only a few more tokens of CPs as responses to the stimuli presented than girls; however, boys produced a broader variety of CP types than girls. 4. Comparing the CPs produced in age group II with those produced in age group I shows that consultants of age group II did not produce the CPs bubwa 'parts cut off, bukwa 'fruit cluster', dumia 'swamplike', and lilo 'walk', which were produced by the consultants of age group I. However, with the consultants of age group II we find the following newly acquired CPs: bogi, bobu/bubu, deli, giwi, gula, gum, kabisi, kabulo, kila, kubila, kudu 'tooth', liku, megwa, mmwa, nunu, po'ula, sipu, suya, tarn, wela, vili, and yam. 5. The CPs iga, kada/keda, ke, kova, kumla, kwe, nal, na2, sisi, ta, and to/te were produced by all consultants in age group II. The CPs uta, ya, and gudi were produced by almost all consultants. The CPs kai, ka'i, and pwanina were produced by all the male consultants, whereas the female consultants were still in the process of acquiring these formatives. The acquisition process of the CPs buda and kwoila had almost been completed in both girls and boys in this age group. Boys acquire the CPs kasa, kweya, giwi, notu, duli, sipu, gum, pila, kubila, liku, and tam earlier or slightly earlier than girls. The acquisition of all the other CPs produced is still in progress. 6. Some of the CPs produced did not agree with the nouns or noun phrases they referred to, according to adult linguistic usage. In addition to these cases of dubious or not acceptable CP usage, there were also quite a few CPs used in ways that showed linguistic insecurity in the consultants with respect to using the CP to refer to a given nominal denotatum appropriately.
52
Classificatory Particles in Kilivila
3.2.1.3 Age Group III Table 3.2.15 gives the number of responses (CP tokens) to the eliciting stimuli presented for demonstrative pronouns, numerals, and adjectives as well as the number of CP types produced by each consultant in age group III (15-20 years). The number of responses elicited from the consultants ranged from 191 (consultant 32) to 207 (consultants 27, 28, 31, and 33-36). Table 3.2.15 records a balanced usage of CPs with adjectives, numerals, and demonstrative pronouns. The consultants of this age group produced 23.7% of all demonstrative pronouns, 23% of all numerals, and 22.7% of all adjectives elicited from the 60 consultants in the study. The male consultants produced only two more CP tokens than the female consultants; however, it should be noted that consultant 32 (male) produced only 191 CP tokens, compared with consultant 25 (female), who produced 198 CP tokens (the lowest number for the six female consultants of this age group). Thus, I conclude from Table 3.2.15 that there are no differences between male and female consultants with respect to the production of CP tokens. The average frequency of CP tokens produced by the female consultants (demonstrative pronouns = 68; numerals = 68; adjectives = 68) differs insignificantly from the average frequency of CP tokens produced by the male consultants (demonstrative pronouns = 68.2; numerals = 68; adjectives = 68.2). The number of types produced by the female consultants ranged from 31 (consultant 25) to 48 (consultant 27); an average of 37.7 CP types were
Table 3.2.15. Responses in age group III Consultant no. (age)
Demonstrative pronoun
Numeral
66 68 69 69 68 68
66 68 69 69 68 68
66 68 69 69 68 68
408
408
408
69 64 69 69 69 69
69 63 69 69 69 69
69
Total
409
408
409
Total, age group III
817
816
817
Females 25 (16y)
26(17y) 27 (20 y) 28(18y) 29(16y) 30(19y) Total Males
31 (19 y)
32(20y) 33(16y) 34(15y) 35 (16y)
36(15y)
Adjective
64
69 69 69 69
No. of types 31 42
48 40
30 35
47 35 37
43 46 35
53
The System of Classificatory Particles in Kilivila
produced. For the male consultants, the number of CP types produced ranged from 35 (consultants 32 and 36) to 47 (consultant 31); an average of 40.5 CP types were produced. Thus, the male consultants produced a slightly broader variety of CP types than the female consultants. Table 3.2.16 lists the CP types that were produced by the consultants of age group III. The CPs gudi, kada, kai, kasa, ke, kova, kumla, kwe, kewya, nal, na2, pwanina, sisi, ta, to/te, and utu were produced by all consultants. Table 3.2.16. CP types and tokens produced in age group III Type beku 'stone blade' bogi 'night' (*) boma '???' bubu 'cut across' bubwa 'part cut off buda 'group' bwa 'tree' bwalita 'sea' deli 'group' duli 'cluster' duya 'door' gili 'row' giwi 'cut' gudi 'child' gula 'heap' gum 'bit' iga 'name' kabisi 'compartment' kabulu 'sector' kada 'road' kaduyo 'entrance' kai 'stone blade' kai 'tooth' kapwa 'parcel' kasa 'row' ke 'wood' kova 'fire' kubila 'land plot' kudu 'tooth' kumla 'earth oven' kwe 'thing'
No. of girls/boys producing CP type
Tokens produced by girls/boys
1/0
1/0 4/4
2/2 1/0
1/0
3/5 1/2
9/14
6/5 4/2 1/1
2/6 25/27 7/5 1/1
6/5 6/5
16/14
0/1 0/1
0/3
5/6 6/6
15/18 18/18
1/4 4/6
12/18
18/15
0/3
3/12
5/6 5/4
14/15
3/2
7/3 44/49 6/2
6/6 1/1 6/6 5/5 3/0 6/6 6/6 6/6 4/6 2/2 6/6 6/6
12/11
17/18 14/14
6/0 28/18
202/178 18/18 9/17
4/4 18/18
292/261
54
Classificatory Particles in Kilivila
Table 3.2.16. (continued) Type kweya limb' kwoila 'clay pot' kwoya/koya 'mountain' liku 'compartment' luba 'bundle' lola 'walk' mmwa 'bundle' megwa 'magic' nal 'female' na2 'animals'
No. of girls/boys producing CP type
Tokens produced by girls/boys
6/6 1/4 2/4 4/4
24/36
3/5 1/0 0/3 2/1 6/6
3/11 4/7 10/12 6/15 3/0 0/8 2/3 21/18
30/24
notulnutu 'kneaded, dot'
6/6 5/6
nunu 'corner/garden'
2/1
peta 'basket'
1/1 5/6 3/3 6/6
3/3 2/1
pila 'part' po'ula 'grove' pwaninalpona 'hole' sa 'nut bunch' si 'small bit' sipu 'sheaf sisi 'bough' suya 'batch of fish' ta 'basket' tam 'sprouting' tuta 'time' tolte 'male' utu 'scrap' uva 'span' vakala 'belt' vaya 'door' vili 'untwisted' vosi 'song' wela 'batch of fish' ya 'thin' yam 'day' yegila 'name' yulai 'bundle of four things' yuva 'shoal' 0 'basketful of yams'
0/1 2/1 4/5 6/6 3/5 6/6 2/2 1/1 6/6 6/6 0/3 1/0 0/1 4/5 2/2 2/2 6/4 2/1 2/2 0/2 1/0 0/1
15/18
24/40 9/9 18/18 0/3 4/1 10/15 25/32
9/15 16/17 9/8 3/1
18/18 66/48 0/9 1/0 0/1 11/15 2/4 5/9 33/10 8/3 4/3 0/6 3/0 0/1
The System of Classificatory Particles in Kilivila
55
Giwi, notu, iga, and pila were produced by 5 female and all male consultants; buda, deli, and duli were produced by all female and 5 male consultants; 5 female and 5 male consultants produced the CP ka'i. All male consultants and 4 female consultants produced the CPs gum and kubila; all female consultants and 4 male consultants produced the CP ya; 5 female consultants and 4 male consultants produced the CP kabisi; 4 females and 5 males produced the CPs sipu and vili. The CP liku was produced by 4 female and 4 male consultants; the CP po'ula was realized by 3 female and 3 male consultants. Five male consultants and 3 female consultants produced the CPs bubu, luba, and suya; 4 male consultants and 2 female consultants produced kwoya/koya, and 4 males and 1 female produced gula and kwoila; 3 female and 2 male consultants produced kabulu. It is feasible to interpret some of these findings in the light of female and male preferences in producing certain CPs. This interpretation may be partly supported by the fact that we find the production of beku, kapwa, lola, vakala, and yuva only by female consultants, and duya, gili, mmwa, sa, uva, vaya, yulai, and 0 (zero classifier) only by male consultants. These findings obviously reflect the slightly broader variety of CP types noted earlier in male consultants compared with female consultants. Table 3.2.17 presents the CP types ordered according to the number of tokens produced by all 12 consultants. Table 3.2.18 lists the expected CPs Table 3.2.17. Number of tokens produced for each CP type—age group III Type kwe 'thing' utu 'scrap' pila 'part' sisi 'bough' buda 'group' va 'thin' gudi 'child' kumla 'earth oven' tolte 'male' duli 'cluster' notulnutu 'kneaded' deli 'group' iga 'name' kubila 'land plot' sipu 'sheaf bubu 'cut across' liku 'compartment' po'ula 'grove'
Tokens 553
114 64 57 52 43
36 36
36 33 33 30 26 26 25 23 22 18
Type ke 'wood' kada 'road' kweya 'limb' na2 'animals' kasa 'row' nal 'female' kova 'fire' pwaninalpona 'hole' kai 'stone blade' giwi 'cut' to 'basket' gum 'bit' vili 'untwisted' vili 'untwisted' suya 'batch/fish' kabisi 'compartment' luba 'bundle' tarn 'sprouting'
Tokens 380 93 60
54 46 39
36 36 35
33 33 30 26 26 24 23 21 17
56
Classificatory Particles in Kilivila
Table 3.2.17. (continued) Type gula 'heap'
Tokens
kwoya/koya 'hill'
15 14 11
kabulu 'sector'
10
wela 'batch/fish'
Type
Tokens
kwoila 'clay pot'
14
bwa 'tree'
12 11
yam 'day' uva 'span'
9
bogi 'night'
8
bubwa 'cut off
8
kaduyo 'entrance'
8 8 6
kudu 'tooth'
8 7
mmwa 'bundle' kapwa 'parcel' vosi 'song'
yegila 'name' nunu 'corner/garden'
6 6 5
yulai 'bundle of four'
tuta 'time'
6 5 4
gili 'row'
3
lola 'walk'
peta 'basket'
3 3
sa 'nut bunch'
3
bwalita 'sea' (*) boma '???'
vakala 'belt'
1 1
2 1
voya 'door'
1
0 'basketful of yams'
1
megwa 'magic'
yuva 'shoal' beku 'stone blade'
si 'bit' duya 'door'
Types total: 68 (69); tokens total: 2,449 (2,450)
Table 3.2. 18. CPs produced versus CPs expected— age group III Produced
Expected
beku 'stone blade'
kai 'stone blade'
bogi 'night'
bogi 'night'
(*) boma '???'
utu 'scrap'
bubu 'cut across'
bubu 'cut across'
bubwa 'cut off'
bubwa 'cut off, kabulo 'point', wtu 'scrap'
budalboda 'group'
buda 'group', bubu 'cut across', yuva 'shoal'
bwa 'tree, wood' bwalita 'sea'
siwa 'sea portions'
deli 'group'
deli 'group'
duli 'cluster'
duli 'cluster'
duya 'door'
duya 'door'
gili 'row'
gili 'row'
giwi 'cut'
giwi 'cut'
bubwa 'cut off, kabulo 'point', utu 'scrap'
gudi 'child'
gudi 'child'
gula 'heap'
gula 'heap', bubu 'cut across', po'ula 'grove'
gum 'bit'
gum 'bit'
3 3
The System of Classificatory Particles in Kilivila
57
Table 3.2.18. (continued) Produced
Expected
iga name kabisi 'compartment' kabulu 'sector'
iga name kabisi 'compartment/foodhouse'
kada 'road' kaduyo 'entrance' kai 'stone blade' ka'i 'tooth' kapwa 'parcel' kasa 'row' ke 'wood'
kova 'fire' koya/kwoya 'hill' kubila 'land plot' kudu 'tooth' kumla 'earth oven' kwe 'thing'
kweya 'limb' kwoila 'clay pot' liku 'compartment' luba 'bundle' lola 'walk' megwa 'magic' mmwa 'bundle' nal 'female' na2 'animals' notulnutu 'kneaded'
kabulu 'village sector' kada 'road', duya 'door', kabulu 'village sector', kaduyo 'entrance' kaduyo 'entrance', duya 'door' kai 'stone blade' kudu 'tooth' kapwa 'parcel' kasa 'row', gili 'row', meila 'song/part', nina 'song/part', deli 'group' ke 'wood', bubu 'cut across', bubwa 'cut off, bukwa 'fruit cluster', bwa 'tree', giwi 'cut', kabisi 'compartment of a foodhouse', kabulo 'sector', kabulo 'point, protuberances' kila 'hands of bananas', kwoila 'clay pot', liku 'compartment of a foodhouse', lila 'bough', luva 'wooden dishes', mmwa 'conical bundle', oyla 'string', sa 'nut bunch', utu 'scrap', uva 'span', ya 'flexible, thin' kova 'fire' kwoya 'mountain, hill' kubila 'land plot' kudu 'tooth' kumla 'earth oven' kwe 'thing', bogi 'night', bubu 'cut across', buda 'group', bukwa 'cluster', deli 'group', duya 'door', gili 'row', gula 'heap', kabisi 'compartment of a foodhouse', kabulu 'village sector', kala 'day', kapwa 'parcel', kubila 'land plot', kudu 'band of fibers', kwoila 'clay pot', kwoya 'mountain', liku 'compartment of a foodhouse', lilo 'walk', nina 'song/part', nunu 'corner of a garden', oyla 'string', po'ula 'grove', sipu 'sheaf, siva 'time', siwa 'sea portions', vilo 'place', yam 'day', yulai 'bundle of four things', yuma 'hand', yuva 'shoal', 0 'basketful of yams' kweya 'limb', yuma 'hand' kwoila 'clay pot' liku 'compartment of a foodhouse' kapwa 'parcel, bundle' lilo 'walk' nina 'song/part' mmwa 'conical bundle' nal 'female persons', duya 'door' na2 'animals', oyla 'string', wela 'batch of fish' notulnutu 'kneaded, dot'
58
Classificatory Particles in Kilivila
Table 3.2.18. (continued) Produced
Expected
nunu corner peta 'basket' pila 'part' po'ula 'grove' pwaninalpona 'hole' sa 'nut bunch' si 'bit' sipu 'sheaf sisi 'bough' suya 'batch of fish' ta 'basket' tam 'sprouting (yams)' tuta 'time' tolte 'male' utu 'scrap'
uva span vakala 'belt' vaya 'door'
nunu 'corner of a garden' ta 'basket' pila 'part', meila 'song/part', nina 'song/part', oyla 'string' po'ula 'grove' pwaninalpona 'hole' sa 'nut bunch' ya 'thin' sipu 'sheaf sisi 'bough', ya 'flexible, thin' wela 'batch of fish' to 'basket' gula 'heap', 0 'basketful of yams' siva 'time' tolte 'male persons' utu 'scrap', kabisi 'compartment/foodhouse', kabulo 'point', kubila 'land plot', meila 'song/part', mmwa 'conical bundle', nina 'song/part', nunu 'corner of a garden', si 'bit', siwa 'sea portions', vili 'untwisted' uva 'span' gili 'row' duya 'door'
vili 'untwisted' vosi 'song' wela 'batch of fish'
vili 'untwisted', si 'bit' meila 'song/part' wela 'batch of fish', oyla 'string'
ya 'thin, flexible' yam 'day' yegila 'name' yulai 'bundle' yuva 'shoal' 0 'basketful of yams'
ya 'flexible, thin', gili 'row', lilo 'walk', pila 'part', sisi 'bough' yam 'day', kala 'day' iga 'name' yulai 'bundle of four things' yuva 'shoal' 0 'basketful of yams'
versus the CPs actually produced by the consultants. The data in this table will be discussed in detail in section 3.3. However, at this stage, the following comments are pertinent. 1. My other consultants told me that there is no CP 'boma' in Kilivila. I have no idea why consultant 25 produced this particle as part of the elicited adjective. 2. Bubwa 'parts cut off was produced as a response to the stimulus presented to elicit the expected CP kabulo 'protuberances'. The CP
The System of Classificatory Particles in Kilivila
59
produced hardly agrees with the stimulus presented; however, if we assume that the consultant thought of a protruding piece of wood that is cut off, we may explain the rather idiosyncratic usage of this CP. 3. Buda/boda 'group, team, crowd' was produced as a response to the stimulus bobu 'log, round timber' presented to elicit the expected CP bubu 'block cut off'. This is not acceptable; the CP buda does not agree with the noun bobu. 4. Gula 'heap, group' was produced as a response to the stimulus bobu 'log, round timber' presented to elicit the expected CP bubu 'block cut off/cut across'. The CP hardly agrees with the stimulus presented; only by assuming that the consultant thought of a heap of blocks cut off can we explain the rather idiosyncratic usage of this CP. 5. Kada 'road, track' was produced as a response to the stimulus kabuluyuvela 'village sector' presented to elicit the expected CP kabulu 'village sector'. This is not acceptable; kada does not agree with the stimulus kabuluyuvela. 6. Nal 'female persons' was produced as a response to the stimulus presented to elicit the expected CP duya 'door, entrance'. This is not acceptable; nal does not agree with the eliciting noun duyava 'door, entrance, window'. However, it may be that the informant tried to play a trick on me by using this CP to give the eliciting noun an obscene meaning. Nevertheless, all my other consultants agreed that this response is not acceptable. 7. Pila 'part, piece' was produced as a response to the stimulus wela 'string of fish' presented to elicit the expected CP oyla 'string'; this is only acceptable if we assume that the consultant thought of a part or a piece of a batch of fish on strings. 8. Utu 'scrap, parts cut off was produced as a response to the stimuli presented to elicit the expected CPs kubila 'land plot', mmwa 'conical bundle', and siwa 'sea portions'. This is not acceptable; utu does not agree with the nouns and phrases presented to elicit the expected CPs. We might justify the use of utu instead of siwa if we assume that the consultant wanted to refer to a part of the sea; nevertheless, the use of utu in this context is very idiosyncratic. Relating these observations to the individual consultants produces the following picture: 5 dubious cases of CP production occurred with consultant 25; consultants 27 and 31 each produced only 1 dubious case of CP production; consultant 30 produced 9 dubious cases; and consultant 32 produced 3 dubious cases. These findings confirm the interpretation of Table 3.2.15 with respect to the balanced usage of CPs with the three word classes elicited in the test. It is also clear that some consultants in age group III made mistakes in the production of CPs and showed linguistic insecurity with respect to the
60
Classificatory Particles in Kilivila
appropriate use of some CPs. This linguistic insecurity becomes even more evident if we look at the cases where individual consultants produced more than one CP type as a response to a single stimulus. Table 3.2.19 presents the responses documenting such multiple classification. This table will be discussed in section 3.3. Table 3.2.20 lists the consultant (see appendix A) and the number of CP types and tokens for demonstrative pronouns, numerals, adjectives, and interrogative pronouns produced in the overall corpus of Kilivila speech data by a speaker of the same age as the consultants in age group III in the elicitation test. Consultant III1 produced these CP tokens during an interview about features of his village and in the context of a public speech he delivered. Table 3.2.21 lists the types and the tokens he produced. The database in this case is obviously too small to provide an analysis. However, I note only that the consultant used the CP kwe to refer to the foreign words Papua New Guinea and kwestin 'question', and that he used the CP pa/pila to refer to the loanword paspoti 'passport'. In summary, the following can be observed with respect to the production of CPs by the consultants in age group III: 1. The data gathered in the CP elicitation test showed a balanced usage of CP tokens with adjectives, numerals, and demonstrative pronouns. There were no differences between male and female consultants with respect to the production of CP tokens; however, the male consultants Table 3.2.19. Responses documenting multiple classification in age group III Cases of multiple classification
Percent
Females 25 26
3 18
10 43
27
9
8 2
20 7
30
9
26
Consultant no.
28 29
19
(Average: 21)
Males 31 32 33
10 1 5
35
5
34
36
21 3 14
8
19
2
6
11
(Average: 12) No. of cases of multiple classification in relation to all CP types produced by the consultant.
The System of Classificatory Particles in Kilivila
61
Table 3.2.20. Consultant and CP types and tokens—overall corpus of Kilivila speech data Consultant no. (age) I I I 1 (
Demonstrative pronoun
Numeral
Adjective
Interrogative pronoun
No. of types
31
21
7
1
5
Table 3.2.21. CP types and tokens produced by the consultant in the overall corpus of Kilivila speech data Type ke 'wood'
Demonstrative pronoun 1 16
Numeral
Adjective
0
0 4 0 0 3 7
tolte 'male'
12
9 0 3 9
Total
31
21
kwe 'thing' pilalpa 'part' na2 'animals'
2 0
Interrogative pronoun
Total
0
1
1 0 0 0
30 2 3 24
1
60
produced a slightly broader variety of CP types than the female consultants did. 2. Consultant 25 produced boma as a response to the stimulus presented to elicit the expected CP utu; however, my other consultants told me that there is no CP boma in Kilivila. The consultants of this age group did not produce the CPs kila 'hands of bananas', luva 'wooden dishes', va 'door', and yuma 'hand', which were produced by the consultants in age group II (see Table 3.2.10). However, the consultants in age group III produced the following CPs, which were not produced by the consultants in age group II: beku, bubwa (this CP was also produced by 1 consultant in age group I, see Table 3.2.2), bwalita, duya, gili, yegila, koya, lola, peta, tuta, uva, vakala, vaya, vosi, yulai, yuva, and 0. 3. The CPs gudi, kada, kai, kasa, ke, kova, kumla, kwe, kweya, nal, na2, pwanina, sisi, ta, to/te, and utu were produced by all consultants; 11 of the 12 consultants produced buda, giwi, iga, notu, pila, deli and duli; 10 consultants produced ka'i, gum, kubila and ya; 9 consultants produced kabisi, sipu, and vili; 8 consultants produced luba, liku, bubu, and suya. With some CPs, we notice gender-specific preferences in this age group. 4. Some of the CPs produced did not agree with the nouns or noun phrases they referred to. In addition to these cases of dubious or not acceptable CP usage, there were also quite a few CPs used in ways that showed linguistic insecurity in the consultants with respect to using the CP to refer to a given nominal denotatum appropriately.
62
Classificatory Particles in Kilivila
I did not use the CPs produced by the consultants of age group III to draw conclusions about language acquisition. Wolfgang Klein formulated the following rule of thumb, which is generally accepted in psycholinguistic research on language acquisition: "Nach der Pubertat entwickelt sich die Sprachbeherrschung nur mehr wenig, obwohl der LernprozeB in manchen Bereichen—z.B. im Wortschatz—nie endet' (Klein 1984:15), and 'Im allgemeinen kann man davon ausgehen, daB sich nach der Pubertat nur noch wenig andert" (Klein 1984:21; see also pp. 16-16, 21-22, 27, 177 notes 2 and 3; ("Beyond puberty, our command of language shows little progress, though in some areas—the vocabulary, for instance—learning continues throughout our life span" [Klein 1986:3] . . . "In general it is reasonable to assume that little progress is made after the age of puberty" [Klein 1986:9; see also pp. 3-4, 8-10, 14-15]); see also Lenneberg 1977:177, 186, 196-197, 202, 208, 220-221, 459, 473; for an extreme counterview, see Schmidt 1975:110). Although this is a cautious approach, I believe it justifies the methods used in the data analyses I will present later in this chapter. To answer questions about acquisition of CPs in Kilivila native speakers, I will use the data produced by the consultants of age group III as one of my primary points of reference (see sections 3.3.2.1-3.3.2.4).
3.2.1.4 Age Group IV Table 3.2.22 gives the number of responses (CP tokens) to the eliciting stimuli presented for demonstrative pronouns, numerals, and adjectives and the number of CP types produced by each consultant in age group IV (21-35 years). The number of responses elicited from the consultants ranged from 204 (consultant 37) to 207 (all other consultants). There was a balanced usage of CPs with adjectives, numerals, and demonstrative pronouns. The consultants of this age group produced 24% of all demonstrative pronouns, 23.4% of all numerals, and 23% of all adjectives elicited from the 60 consultants in the study. With the exception of consultant 37 who produced 3 tokens less than each of the other 11 consultants of this age group, there was a balanced response from all consultants to the stimuli presented with respect to the CP tokens produced. Thus, I infer from the data in Table 3.2.22 that there are no differences between male and female consultants in this age group with respect to the production of CP tokens. The average frequency of CP tokens produced for the female consultants (demonstrative pronouns = 68.8; numerals = 68.8; adjectives = 68.8) differs insignificantly from the average frequency of CP tokens produced for the male consultants (demonstrative pronouns = 69; numerals = 69; adjectives = 69). The number of CP types produced by the female consultants ranged from 18 (consultant 37) to 47 (consultant 41), with an average of 37.2. For male consultants, the number of CP types produced ranged from 31 (consultant 47) to 56 (consultant 45), with an average of 37.5. The difference between
The System of Classificatory Particles in Kilivila
63
Table 3.2.22. Responses in age group IV Consultant no. (age)
Demonstrative pronoun
Numeral
Adjective
No. Of types
Females 37 (26 y) 38 (27 y) 39 (29 y) 40 (28 y) 41 (32 y) 42 (28 y)
68 69 69 69 69 69
68 69 69 69 69 69
68 69 69 69 69 69
18 43 33 44 47 38
413
413
413
69 69 69 69 69 69
69 69 69 69 69 69
69 69 69 69 69 69
Total
414
414
414
Total, age group IV
827
827
827
Total Males 43 (30 y) 44 (35 y) 45 (30 y) 46 (26 y) 47 (25 y) 48 (27 y)
35 32 56 37 31 34
female and male consultants with respect to the production of CP types is insignificant. Table 3.2.23 presents the CP types that were produced by the consultants of age group IV. The CPs kada, kai, ke, kova, kumla, kwe, kweya, nal, na2, sisi, ta, to/te, and utu were produced by all consultants. Giwi, gudi, iga, kasa, pila, pwanina/pona, and suya were produced by 5 female and all male consultants; kubila/kwabila was produced by all female and 5 male consultants; 5 female and 5 male consultants produced the CPs gum, ka'i, kwoila, and ya; 4 female and 5 male consultants produced the CPs duli and vili; 5 female and 4 male consultants produced the CPs kabisi and luba; 4 female and 4 male consultants produced the CPs deli and nutu; 4 female and 3 male consultants produced the CPs liku and sipu; 3 female and 4 male consultants produced the CP gula; 3 female and 3 male consultants produced the CPs buda/boda, tam, and wela; 3 female and 2 male consultants produced the CPs bubu, bubwa, and kudu (tooth); 3 female consultants but only 1 male consultant produced the CP vosi; 3 male consultants but only 1 female consultant produced the CPs bwa and pd'ula. Some of these findings can be interpreted as reflecting female and male preferences in producing certain CPs. This interpretation can be at least partly supported by the fact that the CPs kaduyo, kila, kudu 'fibers', mmwa, nunu, va, and yama were produced only by female consultants, whereas the CPs beku, doba, gili, kabulu, lola, peta, sa, tetu, vakala, and vaya were produced only by male consultants.
64
Classificatory Particles in Kilivila
Table 3.2.23. CP types and tokens produced in age group IV Type beku 'stone blade'
No. of women/men producing CP type
Tokens produced by women/men
0/1
0/1
bogi 'night'
1/1
bubu 'cut across'
3/2
7/6
bubwa 'cut off
3/2 3/3
6/4
budalboda 'group' bwa 'tree'
1/1 16/16
1/6
bwalita 'sea'
1/3 1/1
3/1
deli 'group'
4/4
12/11
doba 'skirt'
0/1
0/1
duli 'cluster'
4/5 0/1
12/15
5/6 5/6 3/4 5/5
13/18
gili 'row' giwi 'cut' gudi 'child' gula 'heap'
gum 'bit' iga 'name' kabisi 'compartment' kabulu 'sector' kada 'road' kaduyo 'entrance'
5/6 5/4 0/2 6/6
0/1 15/18 9/11 15/13 15/18 15/8
0/2 37/41
1/0
kai 'stone blade'
1/0 6/6
18/17
ka'i 'tooth'
5/5
12/14
kapwa 'parcel'
1/1 5/6
16/21
kasa 'row' ke 'wood' kila 'hands/bananas' kova 'fire' kubila 'land plot' kudu 'fibers' kudu 'tooth' kumla 'earth oven' kwe 'thing' kweya 'limb' kwoila 'clay pot' kwoya 'hill' liku 'compartment' lola 'walk' luba 'bundle' luva 'wooden dishes'
3/3
6/6 2/0 6/6 6/5 1/0 3/2
205/207
6/6 6/6 6/6 5/5 2/1 4/3 0/1 5/4 1/2
18/18
4/0 17/18 16/13
3/0 5/4 322/321 33/36 13/13
6/1 12/7
0/1 13/12
4/4
The System of Classificatory Particles in Kilivila
65
Table 3.2.23. (continued) No. of women/men producing CP type
Tokens produced by women/men
mmwa 'conical bundle'
1/0
megwa 'magic'
2/1 6/6 6/6
1/0 5/1 24/27
12/12
Type
mil 'female' na2 'animals'
18/18
9/0
nutu 'kneaded, dot'
2/0 4/4
peta 'basket'
0/1
0/1
pila 'part'
5/6 1/3
33/34
5/6 0/1
13/18 12/9
sisi 'bough'
4/3 6/6
31/33
suya 'batch of fish'
5/6
15/18
to 'basket'
6/6 3/3 0/1
17/17
6/6 1/1
18/18
6/6 2/2
45/51
2/0 0/1 0/1
6/0 0/1 0/2
4/5 3/1
14/11
3/3 5/5
9/9 29/25
1/1 1/0
1/3 2/0
nunu 'corner/garden'
po'ula 'grove' pwanina/pona 'hole' sa 'nut bunch' sipu 'sheaf
tam 'sprouting' tetu 'yams' tolte 'male' tuta 'time' utu 'scrap' uva 'span' va 'door' vakala 'belt' vayalvayo 'door' vili 'untwisted' vosi 'song' wela 'batch of fish' ya 'thin' yam 'day' yama 'hand'
3/8 0/3
14/11
0/1 1/2 2/4
7/3
Table 3.2.24 presents the CP types ordered according to the number of tokens produced by all 12 consultants. Table 3.2.25 lists the CPs produced versus those expected in response to the stimuli presented. The data in this table will be discussed in detail in section 3.3. However, the following should be noted here: 1. The CP bubwa 'block cut across/cut off' was produced as a response to the stimulus presented to elicit the expected CP vili 'untwisted'. This
66
Classificatory Particles in Kilivila
Table 3.2.24. Number of tokens produced for each CP type—age group IV Type
Tokens
kwe 'thing' utu 'scrap' kweya 'limb' sisi 'bough' na2 'animals' kumla 'earth oven' tolte 'male' kova 'fire' gudi 'child' suya 'batch/fish' giwi 'cut' kubila 'land plot' duli 'cluster' kwoila 'clay pot' tam 'sprouting' nutu 'kneaded' kabisi 'compartment' gula 'heap' wela 'batch/fish' po'ula 'grove' vosi 'song' nunu 'corner/garden' bwa 'tree' kapwa 'parcel' uva 'span' bwalita 'sea' yam 'day' sa 'nut bunch' bogi 'night' vaya/vayo 'door' beku 'stone blade' gili 'row' lola 'walk' peta 'basket' vakala 'belt' Types total: 69; tokens total: 2,481
643 96 69 64 51 36 36 35 33 33 31 29 27 26 25 24 23 20 18 11 10 9 7 6 6 4 4 3 2 2 1 1 1 1 1
Type
Tokens
ke 'wood' kada 'road' pila 'part' ya 'flexible' kasa 'row' nal 'female' kai 'stone blade' ta 'basket' iga 'name' budalboda 'group' pwanina/pona 'hole' gum 'bit' kai 'tooth' luba 'bundle' vili 'untwisted' deli 'group' sipu 'sheaf liku 'compartment' bubu 'cut across' bubwa 'cut off kudu 'tooth' luva 'wooden dishes' kwoya 'hill'
412 78
megwa 'magic' va 'door' kila 'hands/bananas' kudu 'fibers' tuta 'time' kabulu 'sector'
6 6 4
yama 'hand' doba 'skirt' kaduyo 'entrance' mmwa 'bundle' tetu 'yams'
2 1 1 1 1
67 54 37 36 35 34 33 32 31 28 26 25 25 23 21 19 13 10 9 8 7
3 3 2
The System of Classificatory Particles in Kilivila
67
Table 3.2.25. CPs produced versus CPs expected—age group IV Produced
Expected
beku 'stone blade' bogi 'night' bubu 'cut across'
kai 'stone blade'
bubwa 'cut off budalboda 'group' bwa 'tree, wood' bwalita 'sea'
bogi 'night' bubu 'block cut across' bubwa 'parts cut off, utu 'scrap', bubu 'cut across', vili 'untwisted' budalboda 'group', gula 'heap', mmwa 'conical bundle', yuva 'shoal' bubu 'cut across', bubwa 'cut off', utu 'scrap' siwa 'sea portions'
deli 'group' doba 'skirt' duli 'cluster'
deli 'group' kudu 'band of fibers' duli 'cluster'
gili 'row' giwi 'cut' gudi 'child'
gili 'row' giwi 'cut'
gula 'heap'
gudi 'child' gula 'heap'
gum 'bit' iga 'name'
gum 'bit' iga 'name'
kabisi 'compartment' kabulu 'sector'
kabisi 'compartment of a foodhouse' kabulu 'village sector'
kada 'road' kaduyo 'entrance'
kada 'road', duya 'door', kaduyo 'entrance' kaduyo 'entrance'
kai 'stone blade' ka'i 'tooth' kapwa 'parcel' kasa 'row'
kai 'stone blade' kudu 'tooth' kapwa 'parcel' kasa 'row', meila 'song/part' ke 'wood', bubu 'cut across', bubwa 'cut off, bukwa 'cluster', bwa 'tree', duya 'door', giwi 'cut', gum 'bit', kabulo 'point', kabulu 'village sector', kabisi 'compartment of a foodhouse', kila 'hands of bananas', kova 'fire', kubila 'land plot', liku 'compartment of a foodhouse', lila 'bough', luva 'wooden dishes', mmwa 'conical bundle', sa 'nut bunch', si 'bit', sisi 'bough', utu 'scrap', uva 'span', vili 'untwisted', ya 'flexible, thin' kila 'hands of bananas' kova 'fire' kubila 'land plot' kudu 'band of fibers' kudu 'tooth'
ke 'wood'
kila 'hands/bananas' kova 'fire' kubila 'land plot' kudu 'fiber' kudu 'tooth'
68
Classificatory Particles in Kilivila
Table 3.2.25. (continued) Produced
Expected
kumla 'earth oven' kwe 'thing'
kumla 'earth oven' kwe 'thing', bogi 'night', bubu 'cut across', bubwa 'cut off, buda 'group', deli 'group', duli 'cluster', duya 'door', gili 'row', giwi 'cut', gula 'heap', gum 'bit', iga name', kabisi 'compartments of a foodhouse', kabulu 'sector', kala 'day', kapwa 'parcel', kasa 'row', kila 'hands of bananas', kubila 'land plot', kudu 'band/fibers', kudu 'tooth', kwoila 'clay pot', kwoya 'mountain', lilo 'walk', meila 'song/part', nina 'song/part', notu 'kneaded, dot', nunu 'corner of a garden', oyla 'string', po'ula 'grove', pwanina 'hole', sa 'nut bunch', sipu 'sheaf, siva 'time', siwa 'sea portions', to 'basket', vili 'untwisted', vilo 'place', yam 'day', yulai 'bundle of four things', yuva 'shoal', 0 'basketful of yams' kweya 'limb', yuma 'hand' kwoila 'clay pot' kwoyalkoya 'mountain, hill' liku 'compartment of a foodhouse' lilo 'walk' kapwa 'parcel, bundle' bukwa 'fruit cluster', mmwa 'conical bundle', sa 'nut bunch'
kweya 'limb' kwoila 'clay pot' kwoya 'hill' liku 'compartment' tola 'walk' luba 'bundle' luva 'wooden dishes' mmwa 'bundle' megwa 'magic' nal 'female' na2 'animals' nunu 'corner/garden' nutulnotu 'kneaded' peta 'basket' pila 'part' po'ula 'grove' pwaninalpona 'hole' sa 'nut bunch' sipu 'sheaf sisi 'bough' suya 'batch of fish' ta 'basket' tarn 'sprouting (yams)' tetu 'yams' tolte 'male' tuta 'time' utu 'scrap'
mmwa 'conical bundle' nina 'part of a song' nal 'female persons' na2 'animals', oyla 'string', wela 'batch of fish' nunu 'corners of a garden' nutulnotu 'kneaded, dot' ta 'basket' pila 'part', meila 'song/part', siva 'time', ya 'flexible, thin', nina 'song/part' po'ula 'grove', sa 'nut bunch' pwaninalpona 'hole' bukwa 'fruit cluster' sipu 'sheaf sisi 'bough', ya 'flexible' wela 'batch of fish' to 'basket' gula 'heap', uva 'span', 0 'basketful of yams' 0 'basketful of yams' tolte 'male persons' kala 'day', siva 'time' utu 'scrap', bubwa 'cut off, kabulo 'point', mmwa 'conical bundle', pila 'part', si 'bit', siva 'time', vili 'untwisted'
The System of Classificatory Particles in Kilivila
69
Table 3.2.25. (continued) Produced
Expected
uva 'span' va 'door' vakala 'belt'
uva 'span' duya 'door' gili 'row'
vayalvayo 'door' vili 'untwisted' vosi 'song' wela 'batch of fish' ya 'thin'
duya 'door', kaduyo 'entrance' vili 'untwisted', gili 'row', pila 'part', si 'bit' meila 'part of a song' oyla 'string' ya 'flexible, thin', bukwa 'fruit cluster', duya 'door', gili 'row', kaduyo 'entrance', pila 'part', sipu 'sheaf, yuma 'hand' yam 'day', kala 'day' yama 'hand'
yam 'day' yama 'hand'
2.
3.
4.
5.
can be explained by assuming that the consultant interpreted the eliciting phrase and gesture in such a way that she emphasized the concept expressed by buwa in her response. Luva 'wooden dishes, tied bundle' was produced as a response to the stimuli presented to elicit the CPs bukwa 'fruit cluster', sa 'nut bunch', and mmwa 'conical bundle'. This can be explained by assuming that the consultants wanted to emphasize the concept 'bunch, bundle' in their responses. Utu 'scrap, parts cut off, fragments' was produced as a response to the stimulus presented to elicit the expected CP mmwa 'conical bundle'. This is not acceptable; utu does not agree with the stimulus budubadu uli 'much taro' presented to elicit the CP mmwa. Vili 'untwisted' was produced as a response to the stimulus presented to elicit the CP gili 'row'. This can be explained by assuming that the consultant wanted to emphasize the concept expressed by vili in connection with the eliciting noun vakala 'belt of spondylus shell discs'; these belts are indeed in safe keeping; they should not be twisted. Ya 'flexible, thin things' was produced as a response to the stimulus presented to elicit the expected CP yam/yama/yuma 'hand'. This may be explained by assuming that the consultant wanted to emphasize the flexibility of the hand; however, this is a rather idiosyncratic usage.
If we relate these observations to the individual consultants, we find the following: With consultants 40 and 48, there was 1 dubious case of CP production; with consultants 42 and 44, there were 3 dubious cases of CP production; and with consultant 41, there were 6 dubious cases of CP production. Thus, some of the consultants showed linguistic insecurity with respect to the approrpiate usage of some CPs. This linguistic insecurity
70
Classificatory Particles in Kilivila
becomes more evident when we look at the cases where individual consultants produced more than one CP type as a response to the stimulus given. Table 3.2.26 presents the responses documenting such multiple classification; it will be discussed in more detail in section 3.3. Table 3.2.27 lists the number of CP types and tokens for demonstrative pronouns, numerals, adjectives, and interrogative pronouns in the overall corpus of Kilivila speech data that were produced by adults of the same age as the consultants in age group IV in the elicitation test (see also appendix A). The data in this table depend heavily on the text category in which the CPs are produced (see appendix B). Consultant IV1/43 produced his CPs in Table 3.2.26. Responses documenting multiple classification in age group IV Cases of multiple classification
Consultant no. Females 37 38 39 40 41 42
Percent
14 7 7 4 9 3
78 16 21 9 19 8
(Average: 25) Males 43 44 45 46 47 48
3 3 22 6 7 3
9 9 39 16 23 9 (Average: 18)
No. of cases of multiple classification in relation to all CP types produced by the consultant.
Table 3.2.27. Consultants and CP types and tokens—overall corpus of Kilivila speech data Consultant no. (age)
Demonstrative pronoun
Numeral
IV 1/43 (30 y) IV2/45 (30 y) IV3 (30 y) IV4 (28 y) IV5 (28 y)
76 5 10 2 1
24 8 23 0 0
9 6 4 0 0
Total
94
55
19
Adjective
Interrogative pronoun
No. of
0 0 3 0 0
10 4 10 2 1
3
types
The System of Classificatory Particles in Kilivila
71
connection with fairy tales, myths, magic, songs, nursery rhymes, and a description of how to carve a tataba-board (which is a sign of honor for chiefs belonging to the Tabulu-subclan); it was in the fairy tales and myths that he produced 37 of the 76 CPs produced with demonstratives. Consultant IV2/45 produced his CPs in connection with a description of how to build a canoe. Consultant IV3 produced his CP tokens in connection with nursery rhymes and descriptions of how to weave a basket and a fishtrap, which accounts for his rather frequent use of numerals. Consultant IV4 produced his CPs in connection with a description of how to make a lime pot, and consultant IV5 produced his CPs in connection with a brief contribution to a public speech. The number of CP types that these 5 men produced ranged from 1 to 10. Table 3.2.28 presents the types and the tokens realized by all 5 consultants. There was no peculiar usage of any of these CP types with respect to the noun they referred to and agreed with. However, the following should be noted: 1. Consultant IV1/43 produced the CP type variant yule 'bundle of four things'; he also produced two tokens of the CP type kwe in its meaning 'clams and shells' to refer to a white cowrie shell (buna). Table 3.2.28. CP types and tokens produced by the consultants in the overall corpus of Kilivila speech data Type
Demonstrative pronoun
gula 'heap'
1
kai 'blade'
0
kasa 'row'
Numeral
Adjective
Interrogative pronoun
Total
1
0 1
0 1
0 0
2
6 4
0 0
0 0
6 4
8 5 4
12 2
1
58 18 4
lipu 'compartment'
0 0 37 11 0 0
0
5
na2 'animals'
20
0
0
pila/pa 'part'
9 1
0 4
1
0 0
2 0 5 2 7 1 1
3 0 0 0 0 0 0
0 0
20 14
0 0 0 2 0 0
18 2 10 1 1
94
55
19
3
171
kauya 'creel' ke 'wood' kwe 'thing' liku 'compartment'
ta 'basket' tam 'sprouting' tolte 'male' uva 'span' ya 'thin' yule 'bundle' yuma 'measure' Total
1 13 0 1
5
0 0
0 0
6 1
72
Classificatory Particles in Kilivila
2. Consultant IV3 produced the CPs kauya 'fish trap', lipu 'compartment of a creel, tier', and yuma 'measure of length, from the fingertips of one hand to the wrist of the other hand'. These CPs were not produced by the consultants in the CP elicitation test, 3. Consultant IV1/43 produced a phrase with double classification: yuletalaga makwena dabunaga yule-tala-ga bundle of four-one-Emphasis
ma-kwe-na this-shell-this
da-buna-ga Dual incl.-cowrie-Emphasis 'indeed one bundle of four of these cowrie shells (belonging to) the two of us' He also used a phrase where he repaired his CP choice: makena . . . e . . . makwena kukwanebu 'this (wooden) . . . eh . . . this (thing) story'. 4. Consultant IV2/45 produced the CP ke to refer with a demonstrative to the loanword riga 'outrigger'. 5. Consultant IV3 produced a phrase where he repaired his CP choice twice: . . . yavila . . . kevila kivaya . . . e . . . yavila . . .' (flexible) how many . . . (wooden) how many fish traps . . . eh (flexible) how many . . .'. To summarize, the following facts can be noted with respect to the production of CPs in age group IV: 1. The data gathered by the CP elicitation test show a balanced usage of CP tokens with adjectives, numerals, and demonstrative pronouns. 2. On average, the differences between male and female consultants with respect to the production of CP tokens and the variety of CP types produced are marginal and insignificant. 3. The consultants in this age group did not produce the CPs duya 'door', yegila 'name', koya 'mountain' (however, they did produce the variant kwoya!), si 'bit', yuva 'shoal', and 0 'basketful of yams' that were produced by the consultants of age group III (see Table 3.2.17). However, the consultants in age group IV produced the following CPs, which were not produced by the consultants in age group III: doba, kauya, kila, kudu 'band of fibers', lipu, luva, tetu, va, yama, and yuma. 4. The CPs kada, kai, ke, kova, kumla, kwe, kweya, nal, na2, sisi, ta, to/te, and utu were produced by all consultants; 11 of the 12 consultants produced the CPs giwi, gudi, iga, kasa, pila, pwanina/pona, suya, and kubila/kwabila', 10 consultants produced the CPs gum, ka'i, kwoila, and ya; 9 consultants produced the CPs duli, kabisi, luba, and vili; 8 consultants produced the CPs deli and nutu; 1 consultants produced the CPs liku, sipu, and gula. With some CPs, gender-specific production preference can be found in the consultants in age group IV.
The System of Classificatory Particles in Kilivila
73
5. There was only one case where the CP produced did not agree with the phrase it referred to; however, with all consultants during the elicitation test we found cases of multiple classification that may document linguistic insecurity. 3.2.1.5 Age Group V Table 3.2.29 gives the number of responses (CP tokens) to the eliciting stimuli presented for demonstrative pronouns, numerals, and adjectives and the number of CP types produced by each consultant in age group V (36-75 years). The number of responses elicited from the consultants ranged from 192 (consultant 52) to 210 (consultant 59). There was a balanced use of CPs with adjectives, numerals, and demonstrative pronouns. The consultants of this age group produced 23.7% of all demonstrative pronouns, 23.1% of all numerals, and 22.7% of all adjectives elicited from the 60 consultants in the study. The male consultants produced only 5 more CP tokens than the female consultants. The average frequency of CP tokens produced by the female consultants (demonstrative pronouns = 67.8; numerals = 68; adjectives = 68) differs insignificantly from the average frequency of CP tokens produced by the male consultants (demonstrative pronouns = 68.2; numerals = 68.3; adjectives = 68.2). Thus, I conclude that there is no significant difference between male and female consultants with respect to the production of CP tokens in this age group. Table 3.2.29. Responses in age group V Consultant no. (age) Females 49 (56 50 (44 51 (48 52 (74 53 (58 54 (39
y) y) y) y) y) y)
Total Males 55 (62 56 (37 57 (45 58 (58 59 (75 60 (48
y) y) y) y) y) y)
Demonstrative pronoun
Numeral
69 68 69 64 68 69
69 68 69 64 69 69
69 68 69 64 69 69
407
408
408
64 69 69 68 70
65
64 69 69 68 70
69 69 68
Adjective
69
70 69
Total
409
410
409
Total, age group V
816
818
817
69
No. Of types
41 32 41 27 36 37
44 37
38 41 37 42
74
Classificatory Particles in Kilivila
The number of CP types produced by the female consultants ranged from 27 (consultant 52) to 41 (consultants 49 and 51), with an average of 35.7. The number of CP types produced by the male consultants ranged from 37 (consultants 56 and 59) to 44 (consultants 55), with an average of 39.8. Table 3.2.30 presents the 72 CP types that were produced by the consultants of age group V. The CPs deli, gudi, iga, kada, kai, kasa, ke, kova, kumla, kwe, kweya, nal, na2, ta, to/te, and utu were produced by all consultants. Notu, Table 3.2.30. CP types and tokens produced by women/men in age group V No. of women/men producing CP type
Tokens produced by women/men 11/14
duli 'cluster'
4/5 3/2 5/5 1/2 6/6 5/5
duya 'door'
0/1
gili 'row'
2/1
5/3
giwi 'cut'
4/5
10/15
guba 'bundle/taro'
1/0
3/0
gudi 'child'
18/18
kabulo 'point'
6/6 2/3 3/4 6/6 2/6 1/0
kabulu 'sector'
2/1
4/3
kada 'road'
36/36
kapwa 'parcel'
6/6 1/1 6/6 5/5 0/1 3/2
kasa 'row'
6/6
18/21
Type bubu 'cut across' bubwa 'cut off' buda/boda 'group' bwa 'tree' deli 'group'
gula 'heap' gum 'bit'
iga 'name' kabisi 'compartment'
kaduyo 'entrance' kai 'stone blade' ka'i 'tooth' kala 'day'
6/4
21/23 2/4
18/18 15/15 0/1
6/7 9/12 18/18 6/16
3/0
1/3
18/18 15/15 0/3 7/4
(*)kava '???'
0/1
0/1
kavi 'tool'
0/1
0/3
ke 'wood'
6/6
219/205
kila 'hands/bananas'
1/0
1/0
kova 'fire' kubila 'land plot'
6/6 3/6
17/18 9/18
kudu 'fibers'
0/2
0/6
kudu 'tooth'
1/1
3/3
The System of Classificatory Particles in Kilivila
75
Table 3.2.30. (continued) Type kumla 'earth oven' kwe 'thing' kweya 'limb' kwottalkwela 'clay pot' kwoya 'mountain' liku 'compartment' lilollola 'walk' luva 'wooden dishes' luba 'bundle' meila 'song/part'
No. of women/men producing CP type
Tokens produced by women/men
6/6 6/6 6/6 3/6 1/0
18/17 310/262
2/5 0/1 1/1 4/4 0/1
30/33 9/6 3/0 6/13 0/2 3/3
11/10 0/3
mmwa 'conical bundle' nal 'female'
2/3
4/8
6/6
na2 'animals' nigwa 'hole'
6/6
18/18 32/36
0/1 0/1
0/1 0/3
notu 'kneaded, dot' nunu 'corner/garden' peta 'basket'
5/6 0/2
13/18 0/6
1/2
pila 'part' po'ula 'grove'
5/5 4/3 5/6
1/2 35/33 12/9
nina 'song/part'
pwaninalpona 'hole' sa 'nut bunch' sipu 'sheaf sisi 'bough'
2/0 5/6 5/6
siva 'time' suya 'batch/fish'
0/1 5/4
ta 'basket' tam 'sprouting' tolte 'male'
6/6 2/1 6/6 1/0 6/6
tuta 'time' utu 'scrap' uva 'span' vili 'untwisted' vosi 'song' wela 'batch/fish' ya 'thin, flexible' yulai 'bundle of four things' yuma 'hand' yuva 'shoal' 0 'basketful of yams'
4/3 5/4 1/0 1/2 3/6 0/1 2/1 0/1 1/0
15/17 6/0 13/18 26/30 0/3 15/12 17/15 14/4 18/20 1/0 30/28
10/6 13/11 2/0 1/6 31/35 0/3 6/1 0/1 1/0
76
Classificatory Particles in Kilivila
pwanina/pona, sipu, and sisi were produced by 5 female and all male consultants; 5 female and 5 male consultants produced the CPs buda/boda, duli, ka'i, and pila; 4 female and 5 male consultants produced the CPs bubu and giwi; 5 female and 4 male consultants produced the CPs suya and vili. All male consultants, but only 3 female consultants produced the CPs kubila, kwoila, and ya. The CP kabisi was produced by all male, but only 2 female consultants; 5 male and 2 female consultants produced the CP liku; 4 female and 4 male consultants produced the CP luba; 4 female and 3 male consultants produced the CPs uva and po'ula; 3 female and 4 male consultants produced the CP gum; 2 female and 3 male consultants produced the CPs gula and mmwa; 3 female and 2 male consultants produced the CPs kapwa and bubwa. Some of these findings seem to reflect female and male preferences in producing certain CPs. This interpretation may be at least partly supported by the fact that only female consultants produced the CPs sa, guba, kabulo, kila, kwoya, tuta, vosi, and 0, and that only male consultants produced the CPs kudu 'band of fibers', nunu, duya, kola, (*)kava, kavi, lilo/lola, meila, nigwa, nina, siva, yulai, yuma, and yuva. It should be noted that my other consultants told me that there is no CP kava in Kilivila. Table 3.2.31 presents the CP types ordered according to the number of tokens produced by all 12 consultants. Table 3.2.32 presents the results of comparing the expected responses to the stimuli presented with the actual responses produced by the consultants. The data presented in this table will be discussed in detail in section 3.3. However, the following comments can be noted here: 1. Kabulo 'village sector' was produced as a response to the stimulus presented to elicit the expected CP nunu 'corners of a garden'. This response can only be explained by assuming that the speaker extended the meaning of kabulo from 'village sector' to 'sector (in general)' and then used the CP in this extended meaning to respond to the stimulus phrase nunula bagula 'corner of a garden'. 2. Consultant 58 produced the formative (*)kava as a response to the noun presented to elicit the expected CP kwoila 'clay pot'; however, all my other consultants told me that there is no CP kava in Kilivila. 3. Kudu 'band of fibers' was produced as a response to the stimuli presented to elicit the expected CPs bubu 'block cut across/cut off and yuva 'shoal'. The CP kudu does not agree with the stimuli presented to elicit these two CPs; thus, these responses are not acceptable. 4. Luva 'wooden dishes, tied bundle' was produced as a response to the stimulus presented to elicit the expected CP utu 'scrap, parts (cut off)'. Luva does not agree with the phrase presented to elicit the expected CP utu; thus, this response is not acceptable. 5. pila 'part, piece' was produced as a response to the stimulus presented to elicit the expected CP oyla 'string'. This response is rather idiosyncratic.
The System of Classificatory Particles in Kilivila
77
Table 3.2.31. Number of tokens produced for each CP type—age group V Type
Tokens
Type
Tokens
56
ke 'wood' na2 'animals' ya 'flexible' utu 'scrap' buda 'group'
39
tolte 'male'
38
36
gudi 'child' km 'tooth' kova 'fire' ta 'basket' notu 'kneaded'
36
duli 'cluster' kubila 'land plot' bubu 'cut across' kwoila 'clay pot' kabisi 'compartment' luba 'bundle' liku 'compartment' uva 'span' mmwa 'bundle' bubwa 'cut off kabulu 'sector'
30 27 25
kwe 'thing' kada 'road'
572 72
pila 'part' kweya 'limb' sisi 'bough' kasa 'row'
68 63
deli 'group' iga 'name' nal 'female' kumla 'earth oven' pwaninalpona 'hole' sipu 'sheaf ka'i 'tooth' suya 'batch/fish' giwi 'cut' vili 'untwisted' gum 'bit' po'ula 'grove' tam 'sprouting' gula 'heap' kapwa 'parcel' gili 'row' wela 'batch/fish' bwa 'tree' kudu 'tooth' nunu 'corner/garden' kaduyo 'entrance' kabulo 'point' kala 'day' meila 'song/part' peta 'basket' yulai 'bundle of four things' vosi 'song' (*)kava '???' nigwa 'hole' yuva 'shoal'
36 36 35 32 31 30 27 25 24 21 21 18 13 11 8 7 6 6 6 4 3 3 3 3 3 2 1 1 1
Types total: 72 (71); tokens total: 2,451 (2,450)
yama 'hand' kudu 'fibers' luva 'wooden dishes' sa 'nut bunch' guba 'bundle/taro' kavi 'tool' kwoya 'mountain' nina 'song/part' siva 'time' lilo 'walk' duya 'door' kila 'hands/bananas' tuta 'time' 0 'basketful/yams'
424 68 66 58 44
36 35 32 31
25 22 21 19 16 12 10 7 7 6 6 6 3 3 3 3 3 2 2 1 1 1
78
Classificatory Particles in Kilivila
Table 3.2.32. CPs produced versus CPs expected—age group V Produced
Expected
bubu 'cut across'
bubu 'block cut across/cut off
bubwa 'cut off
bubwa 'parts cut off', utu 'scrap'
buda/boda 'group'
buda/boda 'group', yuva 'shoal'
bwa 'tree'
bwa 'tree', utu 'scrap', vili 'untwisted'
deli 'group'
deli 'group'
duli 'cluster'
duli 'cluster'
duya 'door'
duya 'door'
gili 'row'
gili 'row'
giwi 'cut'
giwi 'cut'
guba 'bundle/taro'
mmwa 'conical bundle'
gudi 'child'
gudi 'child'
gula 'heap'
gula 'heap'
gum 'bit'
gum 'bit'
iga 'name'
iga 'name'
kabisi 'compartment'
kabisi 'compartment of a foodhouse'
kabulo 'point'
nunu 'corner of a garden'
kabulu 'sector'
kabulu 'village sector'
kada 'road'
kada 'road', duya 'door', kaduyo 'entrance'
kaduyo 'entrance'
kaduyo 'entrance', duya 'door'
kai 'stone blade'
kai 'stone blade'
ka'i 'tooth'
kudu 'tooth'
kala 'day'
kala 'day'
kapwa 'parcel'
kapwa 'parcel'
kasa row'
kasa 'row', meila 'song/part'
(*)kava '???'
kwoila 'clay pot'
kavi 'tool'
kavi 'tool'
ke 'wood'
ke 'wood', bubu 'cut across', bubwa 'cut off, bukwa 'fruit cluster', bwa 'tree', giwi 'cut', gum 'bit', kabisi 'compartment /foodhouse', kabulo 'point', kabulu 'village sector', kila 'hands of bananas', liku 'compartment of a foodhouse', lila 'bough', luva 'wooden dishes', meila 'part of a song', mmwa 'conical bundle', oyla 'string', sa 'nut bunch', si 'bit', sisi 'bough', utu 'scrap', uva 'span', vili 'untwisted', ya 'flexible, thin'
kila 'hands/bananas'
kila 'hands of bananas'
kova 'fire'
kova 'fire'
kubila 'land plot'
kubila 'land plot'
kudu 'fibers'
kudu 'bands of fibers', bubu 'cut across', yuva 'shoal'
kudu 'tooth' kumla 'earth oven'
kudu 'tooth' kumla 'earth oven'
kwe 'thing'
kwe 'thing', bogi 'night', bubu 'cut across', buda 'group', duya 'door', gili 'row', giwi 'cut', gula 'heap', gum 'bit', kabulu
The System of Classificatory Particles in Kilivila
79
Table 3.2.32. (continued) Produced
Expected 'village sector', kaduyo 'entrance', kala 'day', kapwa 'parcel', kova' fire', kubila 'land plot', kudu 'band of fibres', kumla 'earth oven', kwoila 'clay pot', kwoya 'mountain', liku 'compartment of a foodhouse', lilo 'walk', mmwa 'conical bundle', nina 'part of a song', notu 'kneaded', dot', nunu 'corners of a garden', pila 'part', po'ula 'grove', pwanina 'hole', sa 'nut bunch', sipu 'sheaf, siva 'time', siwa 'sea portions', vili 'untwisted', vilo 'place', yam 'day', yulai 'bundle of four things', yuva 'shoal', 0 'basketful of yams'
kweya 'limb'
kweya 'limb', yuma 'hand'
kwoila 'clay pot'
kwoila 'clay pot'
kwoya 'mountain'
kwoya 'mountain'
liku 'compartment'
liku 'compartment of a foodhouse'
lilo/lola 'walk'
lilollola 'walk'
luba 'bundle'
kapwa 'parcel, bundle'
luva 'wooden dishes'
luva 'wooden dishes', utu 'scrap'
meila 'song/part'
meila 'part of a song'
mmwa 'bundle'
mmwa 'conical bundle'
nal 'female'
nal 'female persons'
na2 'animals'
na2 'animals', oyla 'string', wela 'batch of fish'
nigwa 'hole'
pwanina 'hole'
nina 'song/part' notu 'kneaded, dot'
nina 'part of a song' notu 'kneaded, dot'
nunu 'corner/garden'
nunu 'corners of a garden'
peta 'basket'
ta 'basket'
pila 'part'
pila 'part', meila 'part of a song', oyla 'string', nina 'part of a song'
po'ula 'grove'
po'ula 'grove', bukwa 'fruit cluster', sa 'nut bunch'
pwaninalpona 'hole'
pwaninalpona 'hole'
sa 'nut bunch'
sa 'nut bunch'
sipu 'sheaf
sipu 'sheaf
sisi 'bough'
sisi 'bough', ya 'flexible, thin'
siva 'time'
siva 'time'
suya 'batch/fish'
wela 'batch of fish'
ta 'basket'
ta 'basket'
tam 'sprouting (yams)'
gula 'heap', uva 'span', 0 'basketful of yams'
to/te 'male'
to/te 'male persons', yuma 'hand'
tuta 'time'
siva 'time'
utu 'scrap'
utu 'scrap', kabulo 'village sector', si 'bit', siva 'time', vili 'untwisted'
uva 'span'
uva 'span'
vili 'untwisted'
vili 'untwisted', si 'bit'
Classificatory Particles in Kilivila
80
Table 3.2.32. (continued) Produced
Expected
vosi 'song' wela 'batch/fish' ya 'flexible, thin'
meila 'part of a song' wela 'batch of fish', oyla 'string' ya 'flexible, thin', bukwa 'fruit cluster', duya 'door', gili 'row', kaduyo 'entrance', lilo 'walk', pila 'part' yuma 'hand' yulai 'bundle of four things' yuva 'shoal' 0 'basketful of yams'
yama/yuma 'hand' yulai 'bundle of four things' yuva 'shoal' 0 'basketful of yams'
Relating these observations to the individual consultants, we find the following: Consultant 49 produced 3 dubious cases of CP production; consultant 50 produced 1 dubious case; consultant 58 produced 4 dubious cases; and consultant 59 produced 3 dubious cases. Thus, the degree of linguistic insecurity with respect to the appropriate use of CPs is slightly higher in men than in women in this age group. This finding becomes even more striking, if we look at the cases where individual consultants produced more than one CP type as a response to a given stimulus. Table 3.2.33 presents the responses documenting such multiple classification. This table will be discussed in detail in section 3.3. Table 3.2.34 lists the consultants (see appendix A) and the number of CP types and tokens for demonstrative pronouns, numerals, adjectives, and interrogative pronouns in the overall corpus of Kilivila speech data that were produced by adults of the same age as the consultants in age group V in the elicitation test. The data in this table depend heavily on the text category in which the CPs were produced: 1. Consultants Vl/49 and V2/50 produced their CPs in connection with descriptions of how to make a grass skirt; consultant V3/51 produced her CPs in connection with a public speech and a description of how to make a grass-skirt. 2. Consultant V4/53 produced her CPs during an interview about a ghost woman; 10 of the 16 numerals referred to the number of the ghost woman's relatives. 3. Consultant V5/55 produced his CPs during a public speech and in connection with magic formulae; he produced all the adjectives during the recitation of these formulae (see G. Senft 1985c); consultant V6/56 produced his CPs during a description of how to build a house. 4. Consultant V7/60 produced his CPs during an interview about a canoe trip to Nabwageta Island where he bought clay pots. His frequent use of numerals is accounted for by the fact that he referred to the
The System of Classificatory Particles in Kilivila
81
Table 3.2.33. Responses documenting multiple classification in age group V Consultant no.
Cases of multiple classification
Percent
8 13 3 0 1 1
20 41 7 0 3 3
Females 49 50 51 52 53 54
(Average: 12) Males 55 56 57 58 59 60
23 8 8 39 8 10
10 3 3 16 3 4
(Average: 16) No. of cases in relation to all CP types produced by the consultant.
Table 3.2.34. Consultants and CP types and tokens—overall corpus of speech data Consultant no. (age)
Demonstrative pronoun
Numeral
Females Vl/49 V2/50 V3/51 V4/53 V8
(56 y) (44 y) (48 y) (58 y) (60 y)
1 2 3 35 2
0 2 1 16 1
Males V5/55 V6/56 V7/60 V9 V10 V1 1 V12 V1 3 V14 V1 5 V16 V1 7 V1 8
(62 y) (37 y) (48 y) (57 y) (39 y) (63 y) (45 y) (43 y) (60 y) (56 y) (39 y) (41 y) (65 y)
5 1
0
Total
9 1 5 165
4 27
5 20 0 4 59 2 13 13 37
Adjective
l l 5 6 0 14 3 6 0 3 68 0
79
68 17
27 40 6
3 18 5 13 9 3
541
246
158
45 72
Interrogative pronoun
No. Of types
0 0 0 0
2 4
4
7
0
2
0
3 2
0 0 0 3
6 1
0
0
0 0 0 2 0 0 5
4 12 1 5 9
8
9 15 3
Classificatory Particles in Kilivila
82
5.
6.
7.
8.
9.
10.
11.
number of pots he bought and how many pots he intended to give to his relatives and friends. Consultant V8 produced her CPs in connection with songs, consultant V9 produced his CPs during a public ('admonishing') speech (see G. Senft 1991d), and consultant V10 produced his CPs in connection with an admonishing speech, a mythical story, and a prayer. Consultant V11 produced his CPs in connection with a public speech (see G. Senft 1987d), an admonishing speech (see G. Senft 1991d), a mythical story, a joke (see G. Senft 1985e), some magical formulae, and a description of how to build a canoe and carve a canoe prow. Narrating the myth, he produced 71 demonstrative pronouns to refer unequivocally to the protagonists and important objects; describing the canoe prow, he was also forced to refer in an unequivocal way; this explains why he produced so many CPs with demonstrative pronouns. Consultant V12 produced his CPs during an admonishing speech (see G. Senft 1991d). Consultant V13 also produced his CPs during an admonishing speech, and during a description of how to make a sail; both text categories forced him to refer in an unequivocal way, explaining why he produced so many CPs with demonstrative pronouns. Consultant 14 produced his CPs during the narration of an important Trobriand myth; again, the text category required unequivocal references, explaining the rather frequent production of CPs with demonstrative pronouns. Consultant V15 produced his CPs during an interview about a ghost woman (see consultant V4/53) and during a description of how to weave a fish trap and a basket. He produced 28 of the numerals and 64 of the demonstrative pronouns during the interview, where he had to refer to the number of the ghost woman's relatives and where his references to persons and objects had to be unequivocal. Consultant V16 produced his CPs during public speeches and during a description of how to build a house. In the public speeches, he had to refer to a number of persons and objects and was forced to refer in an unequivocal way, explaining the rather frequent production of numerals and demonstrative pronouns. Consultant V17 produced his CPs during a description of how to build a canoe and how to burn chalk; consultant V18 produced his CPs in connection with a public speech and during an admonishing speech (see G. Senft 1991d).
The number of CP types produced by the 18 consultants in this corpus ranged from 1 (consultants V9 and V12) to 15 (consultant V17). Table 3.2.35 presents the types and tokens produced by these consultants. There was no peculiar use of any of these CP types with respect to the nouns they referred
The System of Classificatory Particles in Kilivila
83
Table 3.2.35. CP types/tokens produced by the consultants in the overall corpus of Kilivila speech Type bililo 'trip' bubwa 'cut off buda 'group' bwa 'tree' gudi 'child'
Demonstrative pronoun 0 1 4 0 1
gula 'heap' kabulo 'sector'
3 0
kadalkeda 'road'
3
kai 'stone blade' kali 'paddle strike'
6 0
kasa 'row' kauya 'fish trap' ke 'wood' kova 'fire' kubila 'land plot' kudu 'fibers' kwe 'thing' kwela 'clay pot' kweya 'limb' liku 'compartment nal 'female' na2 'animals' nal 'moon' na4 'carving/human'
Numeral 4 3 7 2
0 0 5 0
2 0
0
0 0 0
4 4 16 2 4 3
1
0 0
2 5
0 1 1
0 0
0 0
6 1
0
0
2 25
0 29
0
1 4
0 0
0
130 4
1
0 0 2
119
82
3 1
I 1
2 91
0 2 1
0 16 0
3
0 0
0 2 75 4 1
25 1
pila 'part' sisi 'bough'
13 1
ta 'basket' tolte 'male'
0 160 0
0 0 2
Total
0 0 0
Total
0
0
yeni 'handful'
1
Interrogative pronoun
2 1
3
utu 'scrap' uva 'span' ya 'thin'
Adjective
82 2
0
5 12 1
541
246
0 21
0 57 4 0
3 0 0 37 0 0 5
1
0 0
1
0
3 258
0 0
8 2
0 0
2 109
0 0 0
26 4 3
0
16 1
0 0 4 0 0 0
2 283 2 5
0
0
38 1
158
5
950
84
Classificatory Particles in Kilivila
to. However, the following facts should be noted: 1. Consultant V6/56 produced a phrase where he repaired his CP choice: ...ketala...kwetala youdila kwetala... '... (wooden)-one... (thing)-one tool (thing)-one... (= one tool)'. 2. Consultant V7/60 produced a phrase with double classification: .. .magulana kwelima kwetala... [infer: kwend]...'...(heap)-this (thing)five (thing)-one... [infer: clay pot]...(= this heap of six clay pots)'. 3. Consultant V11 produced the following two phrases with double classification: ...budakekita tevasi wala... '(group)-small (persons)-four only...(= a small group of four persons only)'; ...tetala vivila e namanabweta minana... '(person)-one girl yes (female)-beautiful (female)this...(= one girl, she was really beautiful)'. He also produced the CP kwe 'thing' to refer with a demonstrative pronoun and an adjective to the foreign word education, to refer with an adjective to the loanword lekodi 'record', to refer with a numeral to the loanword mani 'money', to refer with a demonstrative pronoun to the loan-word laita 'light', and to refer with an adjective to the loanword mitini 'meeting'. Moreover, this consultant produced two tokens of the CP type kwe in its rarely realized meaning 'clams and shells', referring to a triton (tauya). He also produced two phrases where he repaired his CP choice: ...manikwena... mna... mayana vatunu... '(thing)-this...eh...(flexible)-this rope'; ...kweta... budata boda mabudana... '(thing)-one...(group)-one group (group)this... (= this one group)'. Finally, consultant V11 also produced the CP na in its meanings 'moon, stars, planets' ( = na3) and 'carvings in human likeness' ( = na4); here it should be emphasized that Kilagola produced the noun phrase minana tabuya, thus extending the meaning 'carving of human likeness' from the tokwalu-figure on the upper rim of a canoe board (lagim) to the canoe prow (tabuya) (see also Scoditti 1985). 4. Consultant V13 produced the CP ke to refer with a demonstrative pronoun to the foreign word dingi 'dinghy'. Consultant V14 produced the CP kali 'paddle strike'; this CP was not produced by any other consultant. This consultant also produced 15 tokens of the CP type kwe in its meaning 'clams and shells' to refer to the melo aethiopicus shell kweduya. Consultant V15 produced the CP kwe to refer with a demonstrative and with numerals to the loanword tepi 'tape' and the foreign word hour. Consultant 16 produced the CP kwe to refer with demonstratives, numerals, and adjectives to the loanwords ensini 'engine', wiki 'week', namba 'number', simenti 'cement', and mani 'money'; the CP ke to refer with a demonstrative to the foreign word dingi 'dinghy', the CP pila/pa to refer to the loanword paspoti 'passport', and the CP kwela 'claypot' to refer with a numeral to the loanword kerosina 'kerosine'. Moreover, he produced the following phrase where he repaired his CP choice: ...mtona Nina ... minana Nina mtona Gunter...'
The System of Classificatory Particles in Kilivila
85
(person)-this Nina... (female)-this Nina (male)-this Gunter... (this Nina and this Gunter). 5. Informant V17 produced the CP yeni 'a handful of (something)'. This CP, as well as the CPs kali 'paddle strike', kauya 'fish trap', and bililo 'trip', were not produced during the CP elicitation test. Although many of these observations have only anecdotal status, the database seems to be sufficient to confirm the fact that in actual language production the choice of CPs is highly dependent on context, especially on text category. In summary, the following observations apply to the consultants in age group V with respect to their production of CPs: 1. The data gathered by the CP elicitation test showed a balanced usage of CP tokens with adjectives, numerals, and demonstrative pronouns. 2. The differences between male and female consultants with respect to the production of CP tokens are marginal and insignificant; however, on average, the male consultants produced a broader variety of CP types than the female consultants did. 3. The CP sam 'ginger' in phrases like samtala neya 'one piece of ginger' was also produced in actual discourse; however, it was impossible for me to trace the production of this CP to an individual consultant. 4. The consultants of this age group did not produce the CPs beku 'stone blade', bogi 'night', bwalita 'sea', doba 'skirt', lipu 'compartment of a creel', megwa 'magic', tetu 'yams', va 'door', vakala 'belt', vaya/vayo 'door', and yam 'day', which were produced by the consultants of age group IV (see Tables 3.2.23 and 3.2.28). However, the consultants of age group V produced the following CPs, which were not produced by the consultants of age group IV: bililo, duya, guba, kabulo, kala, kali, kavi, meila, na3, na4, nigwa, nina, siva, yeni, yuva, and 0 5. The CPs deli, gudi, iga, kada, kai, kasa, ke, kova, kumla, kwe, kweya, na1, na2, ta, to/te, and utu were produced by all consultants; 11 of the 12 consultants produced the CPs notu, pwanina/pona, sipu, and sisi; 10 consultants produced the CPs buda/boda, duli, ka'i, and pila; 9 consultants produced the CPs bubu, giwi, suya, kubila, kwoila, vili, and ya; 8 consultants produced the CPs kabisi and luba; 1 consultants produced the CPs liku, uva, gum, and po'ula. With some CP types, there were gender-specific production preferences in the consultants of this age group. 6. Some of the CPs produced did not agree with the nouns or noun phrases they referrred to. In addition to these cases of dubious or not acceptable CP usage, quite a few CPs were used in a way that showed linguistic insecurity in some of the consultants.
86
Classificatory Particles in Kilivila
Table 3.2.36. CP types and tokens produced but not assignable to individuals Type bubwa 'cut off buda 'group' kai 'stone blade' kasa 'row' ke 'wood' kwe 'thing' na1 'female' na2 'animals' pila 'part' sam 'ginger' tolte 'male' ya 'flexible' 0 'basketful of yams' Total
Demonstrative pronoun
Numeral
0 2 0
0 0
0 0 13
2 7 11
0 2 6
1
8
1 1 7
0 0 0
7
2
4 3
21 1
28 6
0
2
0
2
20
40
51
111
0
1 8 3 1 0 0 3
1
Adjective
Total 1 2 13 2 10 25 12 2 1
7. The database provided by the overall corpus of Kilivila speech data seems to be sufficient to infer that the choice of CPs in actual language production is highly context dependent Table 3.2.36 lists the CP types and tokens that were produced in actual discourse, but that could not be assigned to individual speakers. These CPs were produced in the following text categories: interview (background), nursery rhymes, verses accompanying games, songs, public speeches, and prayers. The following observations should be noted: The CP sam 'ginger' was not produced during the CP elicitation test. We find the CP to/te produced to refer to a 'corpse', the CP kasa '(a) row (of)' to refer to 'men' and 'girls', and the CP to to refer to the loanword referi 'referee'. 3.2.2 Complete Lists of Data Used in the Analyses So far, I have separately presented and commented on the data produced by the consultants for each of the five age groups. In this section, I list the CP types and the tokens produced by all consultants during the elicitation test; I indicate which consultant produced how many tokens of each CP type. The CP types are listed in alphabetical order. I then present the CP types produced by all consultants during the elicitation test, allocating the tokens
The System of Classificatory Particles in Kilivila
87
of the respective CP type produced to the CP type or types expected. Finally, for the sake of completeness, I list the CP types and tokens documented in my overall corpus of Kilivila speech data. The twofold data presentation given in sections 3.2.1 and 3.2.2, along with the first interpretations and analyses, serves as an empirically sound and checkable basis for the later analyses, which will attempt to answer the questions raised in section 3.1. 3.2.2.1 CP Types/Tokens Produced by All Consultants During the Elicitation Test Tables 3.2.37 to 3.2.123 present the CP types and tokens produced by all consultants during the elicitation test. They indicate which consultant produced how many tokens of each CP type. In the tables, the following abbreviations are used: I-V = age groups; con = consultant no.; f= female; m = male; tok = tokens. Table 3.2.37. Beku 'stone blade' (2 tokens)
I con/tok 1/0
2/0 3/0 4/0 5/0 6/0
con/tok
II
III con/tok
con/tok
IV
V con/tok
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/1 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
43/0 44/0 45/1 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
Total f
1
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
1
Total m Total all con/tok = consultant/token; f = female; m = male.
1
1
88
Classificatory Particles in Kilivila
Table 3.2.38. Bogi 'night' (12 tokens)
I con/tok 1/0 2/0
3/0 4/0 5/0 6/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/1 27/3 28/0 29/0 30/0
37/0 38/1 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
Total f
4
7/0 8/0 9/0
1 43/0 44/0 45/1 46/0 47/0 48/0
19/0 20/0 21/2 22/0 23/0 24/0
31/1 32/0 33/0 34/3 35/0 36/0
Total m
2
4
1
Total all
2
8
2
10/0 11/0 12/0
55/0 56/0 57/0 58/0 59/0 60/0
con/tok = consultant/token; f = female; m = male.
Table 3.2.39. Boma '???' (1 token)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/1 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
1
Total f
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
Total m Total all con/tok = consultant/token; f = female; m = male.
1
The System of Classificatory Particles in Kilivila
89
Table 3.2.40. Bubu/bobu 'block cut off/cut across' (68 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0
13/0 14/0 15/3 16/0 17/0 18/0
25/0 26/3 27/3 28/3 29/0 30/0
37/0 38/3 39/0 40/1 41/3 42/0
49/0 50/2 51/3 52/0 53/3 54/3
3/0 4/0 5/0 6/0
3
9
19/0 20/0 21/0 22/1 23/0 24/3
Total m Total all
Total f
7/0 8/0 9/0 10/0 11/0 12/0
7
11
31/2 32/3 33/0 34/3 35/3 36/3
43/0 44/0 45/3 46/3 47/0 48/0
55/0 56/3 57/3 58/2 59/3 60/3
4
14
6
14
7
23
13
25
con/tok = consultant/token; f = female; m = male.
Table 3.2.41. Bubwa 'parts cut off' (29 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/2 26/0 27/0 28/0 29/0 30/0
37/0 38/2 39/0 40/2 41/2 42/0
49/2 50/1 51/0 52/0 53/0 54/3
2
6
6
31/3 32/0 33/3 34/0 35/0 36/0
43/3 44/0 45/0 46/0 47/1 48/0
55/0 56/3 57/0 58/0 59/0 60/1
Total f
7/0 8/0 9/0 10/0 11/0 12/1
19/0 20/0 21/0 22/0 23/0 24/0
Total m
1
6
4
4
Total all
1
8
10
10
con/tok = consultant/token; f = female; m = male.
90
Classificatory Particles in Kilivila
Table 3.2.42. Buda/boda/budu 'group, team, crowd' (156 tokens)
1
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0
13/0 14/1 15/6 16/1 17/3 18/1
25/1 26/2 27/6 28/4 29/6 30/6
37/0 38/4 39/0 40/6 41/6 42/0
49/3 50/6 51/6 52/0 53/3 54/3
12
25
16
21
10/6 11/2 12/0
19/0 20/0 21/2 22/2 23/1 24/3
31/6 32/6 33/0 34/6 35/6 36/3
43/6 44/0 45/7 46/3 47/0 48/0
55/2 56/6 57/6 58/0 59/3 60/6
Total m
8
8
27
16
23
Total all
8
20
52
32
44
con/tok
3/0 4/0 5/0 6/0 Total f
7/0 8/0 9/0
con/tok = consultant/token; f = female; m = male.
Table 3.2.43. Bukwa 'fruit cluster' (3 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
con/tok 2/0 3/0 4/0 5/0 6/0 Total f
7/0 8/0 9/0 10/3 11/0 12/0 Total m
3
Total all
3
con/tok = consultant/token; f = female; m — male.
The System of Classificatory Particles in Kilivila
91
Table 3.2.44. Bwa 'tree, wood' (61 tokens)
Total f
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/9 5/0 6/0
13/5 14/0 15/5 16/0 17/4 18/0
25/2 26/2 27/2 28/1 29/0 30/0
37/0 38/0 39/1 40/0 41/0 42/0
49/2 50/0 51/0 52/0 53/0 54/0
9
14
7
7/0
19/6 20/5 21/0 22/1 23/0 24/0
31/0 32/0 33/0 34/3 35/2 36/0
8/0 9/0 10/1 11/0 12/0
1 43/0 44/3 45/0 46/1 47/2 48/0
2 55/0 56/0 57/0 58/3 59/0 60/1
Total m
1
12
5
6
4
Total all
10
26
12
7
6
con/tok — consultant/token; f — female; m = male.
Table 3.2.45. Bwalita 'sea' (6 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/1 28/0 29"0 30/0
37/0 38/3 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
1
3
31/0 32/0 33/1 34/0 35/0 36/0
43/0 44/0 45/1 46/0 47/0 48/0
Total f
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
Total m
1
1
Total all
2
4
con/tok = consultant/token; f = female; m = male.
55/0 56/0 57/0 58/0 59/0 60/0
92
Classificatory Particles in Kilivila
Table 3.2.46. Deli 'group on the move' (93 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0
13/0 14/0 15/0 16/0 17/0 18/1
25/3 26/3 27/3 28/1 29/3 30/3
37/0 38/3 39/0 40/3 41/3 42/3
49/3 50/3 51/3 52/3 53/3 54/3
1
16
12
18
31/3 32/3 33/2 34/3 35/3 36/0
43/3 44/0 45/3 46/3 47/2 48/0
55/3 56/3 57/3 58/3 59/3 60/3
con/tok
4/0 5/0 6/0 Total f
7/0
8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/3
Total m
3
14
11
18
Total all
4
30
23
36
con/tok = consultant/token; f = female; m = male.
Table 3.2.47. Doba '(grass) skirt' (1 token) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/1 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
Total f
10/0 11/0 12/0 Total m
1
Total all
1
con/tok = consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
93
Table 3.2.48. Dull 'cluster, bundle' (108 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0
13/0 14/0 15/3 16/0 17/0 18/0
25/3 26/3 27/3 28/3 29/3 30/3
37/0 38/3 39/0 40/3 41/3 42/3
49/3 50/3 51/3 52/0 53/3 54/3
3
18
12
15
10/3 11/0 12/0
19/3 20/3 21/0 22/3 23/0 24/3
31/3 32/0 33/3 34/3 35/3 36/3
43/3 44/0 45/3 46/3 47/3 48/3
55/0 56/3 57/3 58/3 59/3 60/3
Total m
3
12
15
15
15
Total all
3
15
33
27
30
5/0
6/0 Total f
7/0 8/0 9/0
con/tok = consultant/token; f = female; m = male.
Table 3.2.49. Duya 'door, entrance' (4 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
7/0 8/0
19/0 20/0 21/0 22/0 23/0 24/0
31/3 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/1 56/0 57/0 58/0 59/0 60/0
Total f
9/0 10/0 11/0 12/0 Total m
3
1
Total all
3
1
con/tok = consultant/token; f = female; m = male.
94
Classificatory Particles in Kilivila
Table 3.2.50. GIII 'row' (12 tokens)
I con/tok 1/0 2/0 3/0 4/0 5/0 6/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/2 51/0 52/3 53/0 54/0
Total f
5
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/3 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/1 46/0 47/0 48/0
55/3 56/0 57/0 58/0 59/0 60/0
Total m
3
1
3
Total all
3
1
8
con/tok = consultant/token; f = female; m = male.
Table 3.2.51. Giwi 'cut' (106 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0
13/0 14/1 15/3 16/0 17/0 18/0
25/0 26/3 27/3 28/3 29/3 30/3
37/0 38/3 39/1 40/3 41/3 42/3
49/0 50/1 51/3 52/0 53/3 54/3
4
15
13
10
19/3 20/3 21/1 22/3 23/0 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/3 48/3
55/0 56/3 57/3 58/3 59/3 60/3
Total m
13
18
18
15
Total all
17
33
31
25
con/tok 2/0 3/0
4/0
5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0
con/tok — consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
95
Table 3.2.52. Guba 'bundle of taro' (3 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/3 53/0 54/0
Total f
3
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
Total m Total all
3
con/tok = consultant/token; f = female; m = male.
Table 3.2.53. Gudi 'child' (142 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/1 6/0
13/3 14/0 15/3 16/3 17/3 18/3
25/3 26/3 27/3 28/3 29/3 30/3
37/0 38/3 39/3 40/3 41/3 42/3
49/3 50/3 51/3 52/3 53/3 54/3
1
15
18
15
18
7/1 8/0 9/0 10/3 11/1 12/1
19/3 20/3 21/3 22/3 23/0 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/3 48/3
55/3 56/3 57/3 58/3 59/3 60/3
Total m
6
15
18
18
18
Total all
7
30
36
33
36
con/tok
Total f
con/tok = consultant/token; f = female; m = male.
96
Classificatory Particles in Kilivila
Table 3.2.54. Gula 'heap, group' (56 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/5 18/0
25/0 26/0 27/0 28/0 29/3 30/0
37/0 38/0 39/3 40/3 41/3 42/0
49/0 50/0 51/0 52/3 53/3 54/0
5
3
9
6
Total f
7/0 8/0 9/0 10/0 11/0 12/0
19/3 20/0 21/0 22/0 23/0 24/0
31/7 32/0 33/1 34/1 35/3 36/0
43/3 44/3 45/0 46/3 47/2 48/0
55/1 56/0 57/0 58/3 59/3 60/0
Total m
3
12
11
7
Total all
8
15
20
13
con/tok = consultantytoken; f = female; m — male.
Table 3.2.55. Gum 'bit small piece' (88 tokens) II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/1 17/2 18/0
25/0 26/3 27/3 28/3 29/0 30/3
37/0 38/3 39/3 40/3 41/3 42/3
49/0 50/0 51/3 52/0 53/3 54/3
3
12
15
9
19/1 20/1 21/0 22/1 23/0 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/1 48/0
55/0 56/3 57/3 58/0 59/3 60/3
Total m
6
18
13
12
Total all
9
30
28
21
I con/tok 1/0
2/0 3/0
4/0 5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0
con/tok = consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
97
Table 3.2.56. Iga 'name' (142 tokens)
I con/tok
II con/tok
III con/tok
13/3 14/1
1/0 2/0
3/0 4/3 5/0 6/1
15/3 16/3 17/3 18/3
25/3 26/2 27/0 28/3 29/3 30/3
4
16
7/0
IV con/tok
V con/tok
37/0
49/3
41/3 42/3
50/3 51/3 52/3 53/3 54/3
14
15
18
10/3 11/3 12/0
19/3 20/3 21/3 22/3 23/3 24/3
31/3 32/3 33/3 34/3 35/1 36/2
43/3 44/3 45/3 46/3 47/3 48/3
55/3 56/3 57/3 58/3 59/3 60/3
Total m
6
18
15
18
18
Total all
10
34
29
33
36
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/1 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
Total f
8/0 9/0
38/3 39/3 40/3
con/tok = consultant/token; f = female; m = male.
Table 3.2.57. Iki '???' (1 token)
I con/tok 1/0 2/0 3/0 4/0 5/0 6/0
Total f
1
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
Total m Total all
1
con/tok = consultant/token; f = female; m = male.
Classificatory Particles in Kilivila
98
Table 3.2.58. Kabisi 'compartment of a foodhouse' (78 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/3 14/0 15/0 16/0 17/1 18/0
25/3 26/1 27/3 28/0 29/3 30/2
37/0 38/3 39/3 40/3 41/3 42/3
49/3 50/0 51/3 52/0 53/0 54/0
4
12
15
6
19/3 20/0 21/3 22/0 23/0 24/0
31/2 32/0 33/3 34/3 35/3 36/0
43/0 44/1 45/1 46/3 47/0 48/3
55/3 56/3 57/3 58/3 59/3 60/1
Total m
6
11
8
16
Total all
10
23
23
22
con/tok
Total f
7/0 8/0 9/0 10/0 11/0 12/0
con/tok — consultant/token; f = female; m — male.
Table 3.2.59. Kabulo 1 'protuberances, point' (3 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/3 50/0 51/0 52/0 53/0 54/0
con/tok
Total f
3
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
Total m Total all con/tok = consultant/token; f = female; m = male.
3
The System of Classificatory Particles in Kilivila
99
Table 3.2.60. Kabulo2/kabulu Village sector' (21 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/1 27/3 28/3 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/1 50/0 51/0 52/0 53/3 54/0
Total f
7
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/2
31/2 32/0 33/0 34/1 35/0 36/0
4
43/0 44/0 45/1 46/1 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/3
Total m
2
3
2
3
Total all
2
10
2
7
con/tok = consultant/token; f = female; m = male.
Table 3.2.61. Kada/keda 'road, track' (341 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0
13/5 14/8 15/6 16/3 17/3 18/7
25/6 26/8 27/3 28/9 29/9 30/9
37/8 38/3 39/6 40/9 41/5 42/6
49/3 50/6 51/6 52/6 53/6 54/9
con/tok 2/0 3/0 4/0 5/0 6/5
5
32
44
37
36
7/0 8/0 9/0 10/3 11/0 12/9
19/9 20/3 21/9 22/9 23/10 24/9
31/6 32/12 33/7 34/6 35/9 36/9
43/9 44/9 45/5 46/9 47/3 48/6
55/7 56/3 57/6 58/8 59/3 60/9
Total m
12
49
49
41
36
Total all
17
81
93
78
72
Total f
con/tok = consultant/token; f = female; m = male.
100
Classificatory Particles in Kilivila
Table 3.2.62. Kaduyo/kaduya 'door, entrance' (15 tokens) con/tok
con/tok
II
III con/tok
IV con/tok
con/tok
1/0 2/0 3/0 4/1 5/0
6/0
13/1 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/6 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/1 42/0
49/0 50/1 51/0 52/0 53/0 54/0
1
1
6
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
I
Total f
10/0 11/0 12/0
31/0 32/0 33/2 34/0 35/0 36/0
Total m Total all
1
1
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/3 58/0 59/0 60/0
3
2 1
1
8
V
1
4
con/tok = consultant/token; f = female; m = male.
Table 3.2.63. Kal 'stone blade' (136 tokens)
I
II con/tok
III con/tok
con/tok
con/tok
1/0 2/0 3/0 4/0
13/0 14/1 15/3 16/0 17/3 18/3
25/3 26/2 27/3 28/3 29/3 30/3
37/3 38/3 39/3 40/3 41/3 42/3
49/3 50/3 51/3 52/3 53/3 54/3
10
17
18
18
10/3 11/0 12/0
19/3 20/3 21/3 22/3 23/2 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/2 46/3 47/3 48/3
55/3 56/3 57/3 58/3 59/3 60/3
Total m
3
17
18
17
18
Total all
3
27
35
35
36
con/tok
5/0 6/0
Total f 7/0 8/0 9/0
con/tok = consultant/token; f = female; m = male.
IV
V
The System of Classificatory Particles in Kilivila
101
Table 3.2.64. Ka'i 'tooth' (127 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
2/0 3/0 4/2 5/3 6/3
1/0
13/0 14/0 15/0 16/3 17/3 18/3
25/3 26/3 27/3 28/3 29/0 30/2
37/2 38/2 39/3 40/0 41/2 42/3
49/3 50/3 51/3 52/0 53/3 54/3
8
9
14
12
15
11/2 12/0
19/3 20/3 21/3 22/3 23/6 24/3
31/3 32/3 33/3 34/2 35/3 36/0
43/3 44/0 45/3 46/3 47/3 48/2
55/3 56/3 57/3 58/3 59/0 60/3
Total m
5
21
14
14
15
Total all
13
30
28
26
30
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/3 56/0 57/0 58/0 59/0 60/0
Total f
7/0 8/0 9/0 10/3
con/tok = consultant/token; f = female; m = male.
Table 3.2.65. Kala 'day' (3 tokens)
I con/tok 1/0 2/0
3/0 4/0 5/0
6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0 Total m
3
Total all
3
con/tok — consultant/token; f = female; m = male.
Classificatory Particles in Kilivila
102
Table 3.2.66. Kapwa/kapo 'bundle/parcel wrapped up' (46 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0
13/3 14/0 15/3 16/0 17/0 18/3
25/3 26/0 27/0 28/1
37/0 38/0 39/0 40/0 41/0 42/3
49/1 50/3 51/0 52/0 53/0 54/3
9
6
7/0 8/0 9/0 10/2 11/3 12/0
19/3 20/3 21/0 22/3 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
Total m
5
9
Total all
5
18
con/tok
4/0 5/0 6/0 Total f
29/0 30/2
3
43/0 44/0 45/0 46/0
7
55/3 56/0 57/0 58/1
48/0
59/0 60/0
3
4
6
6
11
III
IV con/tok
V con/tok
25/3 26/3 27/4
37/0 38/3
49/3
47/3
con/tok — consultant/token; f = female; m = male.
Table 3.2.67. Kaso 'row, line' (165 tokens)
I con/tok 1/0 2/0
13/0
con/tok
14/0 15/3 16/0 17/3 18/3
28/5 29/10 30/3
40/3 41/3 42/3
54/3
3
9
28
16
18
7/0 8/0 9/0 10/2 11/3
19/4
31/3 32/3 33/3 34/3 35/3 36/3
43/3
55/3
45/3 47/3 48/3
57/3 58/3 59/6 60/2
3/0 4/0
5/0 6/3
Total f
II con/tok
39/4
44/3
50/3 51/3 52/3 53/3
56/3
12/3
20/3 21/3 22/3 23/10 24/0
Total m
8
23
18
21
21
Total all
11
32
46
37
39
con/tok = consultant/token; f = female; m = male.
46/6
The System of Classificatory Particles in Kilivila
103
Table 3.2.68. Kava '???' (1 token)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/1 59/0 60/0
con/tok
3/0 4/0 5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0 Total m
1
Total all
1
con/tok = consultant/token; f = female; m = male.
Table 3.2.69. Kavl 'tool' (3 tokens) I con/tok
If con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/3 60/0
3/0 4/0 5/0 6/0 Total f
10/0 11/0 12/0 Total m
3
Total all
3
con/lok = consultant/token; f = female; m = male.
104
Classiflcatory Particles in Kilivila
Table 3.2.70. Ke 'wood' ('long objects, inanimates, fire') (1,919 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/3 2/0 4/23 5/26 6/33
13/34 14/31 15/35 16/44 17/38 18/50
25/36 26/28 27/28 28/35 29/43 30/32
37/47 38/27 39/28 40/30 41/32 42/41
49/33 50/40 51/29 52/48 53/39 54/30
85
232
202
205
219
7/3 8/3 9/0 10/31 11/36 12/46
19/37 20/40 21/49 22/39 23/54 24/48
31/22 32/39 33/33 34/25 35/22 36/37
43/33 44/35 45/33 46/36 47/38 48/32
55/37 56/33 57/35 58/34 59/36 60/30
Total m
119
267
178
207
205
Total all
204
499
380
412
424
3/0
Total f
con/tok = consultant/token; f = female; m = male.
Table 3.2.71. Kila 'clusters/hands of bananas' (7 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0
13/0 14/0 15/2 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/3 40/0 41/1 42/0
49/1 50/0 51/0 52/0 53/0 54/0
4/0
5/0
6/0
2
Total f
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
4
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
1
55/0 56/0 57/0 58/0 59/0 60/0
Total m Total all
2
con/tok = consultant/token; f = female; m = male.
4
1
The System of Classificatory Particles in Kilivila
105
Table 3.2.72. Kova 'fire' (159 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
2/0 3/0 4/3 5/3
1/0
6/3
13/3 14/3 15/3 16/3 17/3 18/3
25/3 26/3 27/3 28/3 29/3 30/3
37/2 38/3 39/3 40/3 41/3 42/3
49/3 50/2 51/3 52/3 53/3 54/3
9
18
18
17
17
7/0 8/0 9/0 10/3 11/2 12/3
19/3 20/3 21/3 22/3 23/3 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/3 48/3
55/3 56/3 57/3 58/3 59/3 60/3
Total m
8
18
18
18
18
Total all
17
36
36
35
35
con/tok
Total f
con/tok — consultant/token; f = female; m = male.
Table 3.2.73. Kubila/kwabila 'land plot' (94 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0
13/0 14/0 15/0 16/0 17/3 18/0
25/3 26/2 27/3 28/0 29/0 30/1
37/1 38/3 39/3 40/3 41/3 42/3
49/3 50/0 51/3 52/3 53/0 54/0
3
9
16
9
19/3 20/0 21/3 22/0 23/0 24/3
31/2 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/2 46/2 47/0 48/3
55/3 56/3 57/3 58/3 59/3 60/3
Total m
9
17
13
18
Total all
12
26
29
27
con/tok 2/0
3/0 4/0 5/0
6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0
con/tok = consultant/token; f = female; m = male.
106
Classificatory Particles in Kilivila
Table 3.2.74. Kudu1 'band of fibres' (9 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/3
49/0 50/0 51/0 52/0 53/0 54/0
con/tok
Total f
3
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/3 56/0 57/0 58/3 59/0 60/0
6
Total m Total all
3
6
con/tok = consultant/token; f = female; m = male.
Table 3.2.75. Kudu2 'tooth' (26 tokens)
I con/tok 1/0
2/0 3/0 4/0 5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/3 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/3 30/1
37/0 38/1 39/0 40/3 41/1 42/0
49/0 50/0 51/0 52/3 53/0 54/0
3
4
5
3
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/1 35/0 36/3
43/0 44/3 45/0 46/0 47/0 48/1
55/0 56/0 57/0 58/0 59/3 60/0
4
4
3
8
9
6
Total m Total all
3
con/tok = consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
107
Table 3.2.76. Kumla 'earth oven' (162 tokens)
II con/tok
III con/tok
IV con/tok
V con/tok
2/0 3/0 4/3 5/3 6/3
1/1
13/3 14/3 15/3 16/3 17/3 18/3
25/3 26/3 27/3 28/3 29/3 30/3
37/3 38/3 39/3 40/3 41/3 42/3
49/3 50/3 51/3 52/3 53/3 54/3
10
18
18
18
18
7/0 8/0 9/0 10/3 11/3 12/3
19/3 20/3 21/3 22/3 23/3 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/3 48/3
55/2 56/3 57/3 58/3 59/3 60/3
Total m
9
18
18
18
17
Total all
19
36
36
36
35
IV con/tok
V con/tok
I con/tok
Total f
con/tok = consultant/token; f = female; m = male.
Table 3.2.77. Kwe 'thing, inanimates' (2,877 tokens) I
con/tok
1/1
II con/tok
III
con/tok
13/107 14/52 15/48 16/80 17/62 18/89
25/65 26/58 27/27 28/41 29/55 30/46
37/101 38/34 39/69 40/40 41/39 42/39
49/44 50/71 51/35 52/54 53/55 54/51
134
438
292
322
310
7/3
10/51 11/55 12/59
19/59 20/74 21/49 22/74 23/64 24/46
31/36 32/27 33/62 34/46 35/34 36/56
43/42 44/57 45/30 46/54 47/72 48/66
55/33 56/40 57/45 58/47 59/55 60/42
Total m
171
366
261
321
262
Total all
305
804
553
643
572
2/0 3/0 4/39 5/51 6/43
Total f
8/3
9/0
con/tok = consultant/token; f = female; m = male.
108
Classificatory Particles in Kilivila
Table 3.2.78. Kweya/kwaya/keya "(severed) limb' (215 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0
13/0 14/0 15/0 16/3 17/3 18/3
25/3 26/3 27/6 28/6 29/3 30/3
37/6 38/6 39/6 40/3 41/6 42/6
49/6 50/6 51/3 52/6 53/3 54/6
9
24
33
30
10/3 11/0 12/0
19/1 20/3 21/3 22/3 23/1 24/0
31/6 32/6 33/6 34/6 35/6 36/6
43/6 44/6 45/6 46/6 47/6 48/6
55/3 56/6 57/6 58/6 59/6 60/6
Total m
3
11
36
36
33
Total all
3
20
60
69
63
5/0
6/0 Total f
7/0 8/0 9/0
con/tok = consultant/token; f = female; m = male.
Table 3.2.79. Kwoila/kwela 'clay pot' (96 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/3 14/0 15/3 16/3 17/3 18/3
25/0 26/0 27/0 28/3 29/0 30/0
37/3 38/0 39/1 40/3 41/3 42/3
49/3 50/3 51/0 52/0 53/0 54/3
15
3
13
9
10/2 11/0 12/3
19/3 20/2 21/3 22/3 23/0 24/0
31/2 32/3 33/0 34/0 35/3 36/3
43/3 44/3 45/3 46/0 47/1 48/3
55/3 56/3 57/3 58/2 59/2 60/3
Total m
5
11
11
13
16
Total all
5
26
14
26
25
Total f
7/0 8/0 9/0
con/tok — consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
109
Table 3.2.80. Kwoya/koya 'mountain, hill' (26 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
2/0 3/0 4/0 5/0 6/0
l/l
13/0 14/0 15/1 16/0 17/0 18/0
25/0 26/0 27/3 28/1 29/0 30/0
37/0 38/0 39/0 40/0 41/3 42/3
49/0 50/0 51/0 52/0 53/3 54/0
1
1
con/tok
Total f
7/0 8/0 9/0 10/0 11/0 12/0 Total m Total all
1
4
6
19/0 20/0 21/3 22/0 23/0 24/0
31/2 32/2 33/0 34/2 35/1 36/0
3
7
1
4
11
7
43/0 44/0 45/1 46/0 47/0 48/0
3
55/0 56/0 57/0 58/0 59/0 60/0
3
con/tok = consultant/token; f = female; m = male.
Table 3.2.81. Liku 'compartment of a foodhouse' (71 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0
13/0 14/0 15/1 16/0 17/1 18/0
25/0 26/1 27/3 28/3 29/0 30/3
37/0 38/3 39/3 40/3 41/0 42/3
49/3 50/0 51/3 52/0 53/0 54/0
2
10
12
6
19/3 20/0 21/3 22/0 23/0 24/3
31/0 32/3 33/3 34/3 35/3 36/0
43/3 44/0 45/1 46/0 47/0 48/3
55/3 56/3 57/0 58/2 59/3 60/2
Total m
9
12
7
13
Total all
11
22
19
19
3/0 4/0 5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0
con/tok = consultant/token; f = female; m = male.
110
Classificatory Particles in Kilivila
Table 3.2.82. Lilo/lola 'walk, Journey' (7 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/1
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/3 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
43/0 44/0 45/1 46/0 47/0 48/0
55/2 56/0 57/0 58/0 59/0 60/0
con/tok 2/0 3/0 4/0 5/0 6/0 Total f
1 7/0 8/0 9/0 10/0 11/0 12/0
3
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
Total m Total all
1
3
1
2
1
2
con/tok = consultant/token; f = female; m = male.
Table 3.2.83. Luba 'bundle' (77 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/3 18/0
25/0 26/0 27/3 28/2 29/0 30/1
37/1 38/3 39/3 40/3 41/3 42/0
49/2 50/0 51/3 52/3 53/3 54/0
3
6
13
11
10/1 11/0 12/0
19/0 20/0 21/3 22/0 23/0 24/3
31/3 32/3 33/3 34/3 35/3 36/0
43/0 44/3 45/3 46/3 47/0 48/3
55/0 56/2 57/3 58/2 59/0 60/3
Total m
1
6
15
12
10
Total all
1
9
21
25
21
con/tok
Total f
7/0 8/0 9/0
con/tok = consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
111
Table 3.2.84. Luva 'wooden dishes' (18 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0
13/0 14/0 15/0 16/0 17/3 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/4 42/0
49/0 50/0 51/0 52/0 53/0 54/3
4
3
43/0 44/3 45/0 46/0 47/0 48/1
55/0 56/0 57/0 58/0 59/3 60/0
4
3
8
6
4/0
5/0 6/0 Total f
3
7/0 8/1 9/0 10/0 11/0 12/0 Total m
1
Total all
1
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
3
con/tok = consultant/token; f = female; m = male.
Table 3.2.85. Megwa 'magic' (12 tokens)
I con/tok 1/0
2/0 3/0 4/0 5/0 6/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/1 28/1 29/0 30/0
37/0 38/3 39/0 40/2 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
Total f
2
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/1
31/0 32/0 33/3 34/0 35/0 36/0
5
43/0 44/0 45/1 46/0 47/0 48/0
Total m
1
3
1
Total all
1
5
6
con/tok = consultant/token; f = female; m = male.
55/0 56/0 57/0 58/0 59/0 60/0
112
Classificatory Particles in Kilivila
Table 3.2.86. Meila 'part of a song' (3 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 43/0 54/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/3 56/0 57/0 58/0 59/0 60/0
con/tok
5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0 Total m
3
Total all
3
con/tok = consultant/token; f = female; m = male.
Table 3.2.87. Mmwa 'conical bundle' (23 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/1 42/0
49/3 50/0 51/1 52/0 53/0 54/0
1
Total f
7/0 8/0 9/0 10/0 11/0 12/0
19/2 20/0 21/0 22/0 23/0 24/0
31/3 32/0 33/0 34/2 35/3 36/0
Total m
2
8
Total all
2
8
con/tok — consultant/token; f = female; m = male.
43/0 44/0 45/0 46/0 47/0 48/0
4
55/3 56/0 57/3 58/2 59/0 60/0
8 1
12
The System of Classificatory Particles in Kilivila
113
Table 3.2.88. Na1 'persons of female gender' (184 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/3 2/1 3/1 4/4 5/4 6/3
13/6 14/3 15/3 16/3 17/3 18/3
25/6 26/3 27/3 28/3 29/3 30/3
37/3 38/3 39/3 40/3 41/3 42/3
49/3 50/3 51/3 52/3 53/3 54/3
16
21
21
18
18
7/3 8/3 9/1 10/3 11/5 12/3
19/3 20/3 21/3 22/3 23/3 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/3 48/3
55/3 56/3 57/3 58/3 59/3 60/3
Total m
18
18
18
18
18
Total all
34
39
39
36
36
Total f
con/tok = consultant/token; f = female; m = male.
Table 3.2.89. Na2 'animals' (279 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/6 5/9 6/6
13/4 14/6 15/6 16/6 17/3 18/3
25/3 26/6 27/3 28/9 29/3 30/6
37/6 38/3 39/3 40/3 41/3 42/6
49/4 50/7 51/6 52/3 53/6 54/6
21
28
30
24
32
7/0 8/0 9/0 10/6 11/6 12/6
19/6 20/6 21/6 22/9 23/6 24/6
31/3 32/3 33/3 34/6 35/6 36/3
43/6 44/6 45/3 46/3 47/6 48/3
55/6 56/6 57/6 58/6 59/9 60/3
Total m
18
39
24
27
36
Total all
39
67
54
51
68
Total f
con/tok = consultant/token; f = female; m = male.
114
Classiflcatory Particles in Kilivila
Table 3.2.90. Nigwa 'hole' (1 token)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/1 56/0 57/0 58/0 59/0 60/0
con/tok 2/0 3/0 4/0 5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0 Total m
1
Total all
1
con/tok = consultant/token; f = female; m = male.
Table 3.2.91. Nina 'part of a song' (3 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/3 56/0 57/0 58/0 59/0 60/0
Total f
10/0 11/0 12/0 Total m
3
Total all
3
con/tok = consultant/token; f = female; rn = male.
The System of Classificatory Particles in Kilivila
115
Table 3.2.92. Notu/nutu 'kneaded, dot, drop' (103 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/3 18/0
25/0 26/3 27/3 28/3 29/3 30/3
37/0 38/3 39/0 40/3 41/3 42/3
49/3 50/1 51/3 52/0 53/3 54/3
3
15
12
13
10/0 11/1 12/2
19/1 20/1 21/3 22/1 23/0 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/0 48/0
55/3 56/3 57/3 58/3 59/3 60/3
Total m
3
9
18
12
18
Total all
3
12
33
24
31
1/0 2/0
3/0 4/0 5/0 6/0 Total f
7/0 8/0
9/0
con/tok = consultant/token; f = female; m = male.
Table 3.2.93. Nunu 'corner of a garden' (22 tokens)
I con/tok 1/0 2/0 3/0 4/0 5/0 6/0 Total f
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/1 16/0 17/0 18/0
25/0 26/2 27/0 28/1 29/0 30/0
37/0 38/3 39/3 40/0 41/0 42/3
49/0 50/0 51/0 52/0 53/0 54/0
1
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/3 36/0
9
43/0 44/0 45/0 46/0 47/0 48/0
3
Total m Total all
3
1
con/tok = consultant/token; f = female; m = male.
6
55/3 56/3 57/0 58/0 59/0 60/0
6 9
6
116
Classificatory Particles in Kilivila
Table 3.2.94. Peta 'basket' (7 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/2
37/0 38/0
49/0 50/1 51/0 52/0 53/0 54/0
28/0 29/0 30/0
39/0 40/0 41/0 42/0
2
Total f
1 55/0
35/0 36/1
43/0 44/0 45/1 46/0 47/0 48/0
Total m
1
1
2
Total all
3
1
3
8/0 9/0
19/0
31/0
20/0 21/0
10/0 11/0 12/0
22/0 23/0 24/0
32/0 33/0 34/0
7/0
56/0 57/1 58/0 59/0
60/1
con/tok = consultant/token; f = female; m = male.
Table 3.2.95. Pila/po 'part piece' (249 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/3 14/0 15/8 16/0 17/6 18/0
25/6 26/5
49/6
27/7 28/3 29/0 30/3
37/4 38/0 39/8 40/3 41/8 42/10
17
24
33
35
19/6
10/7 11/1 12/3
20/3 21/4 22/9 23/0 24/0
31/8 32/9 33/1 34/8 35/8 36/6
43/7 44/10 45/2 46/3 47/3 48/9
55/0 56/4 57/6 58/9 59/5 60/9
Total m
11
22
40
34
33
Total all
11
39
64
67
68
Total f 7/0 8/0 9/0
con/tok = consultant/token; f = female; m = male.
50/7 51/7 52/9 53/0 54/6
The System of Classificatory Particles in Kilivila
117
Table 3.2.96. Po'ula 'plantation, grove' (53 tokens)
I con/tok 1/0
2/0 3/0
4/0 5/0 6/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/3 28/3 29/3 30/0
37/0 38/0 39/0 40/3 41/0 42/0
49/3 50/0 51/3 52/0 53/3 54/3
9
3
12
Total f
7/0 8/0 9/0
19/0 20/0 21/3 22/0 23/0 24/0
31/0 32/3 33/0 34/0 35/3 36/3
Total m
3
9
8
9
Total all
3
18
11
21
10/0 11/0 12/0
43/0 44/0 45/3 46/3 47/0 48/2
55/0 56/0 57/3 58/3 59/0 60/3
con/tok = consultant/token; f = female; m = male.
Table 3.2.97. Pwonina/pono 'punctured, hole' (140 tokens)* I con/tok
con/tok
II
III con/tok
IV con/tok
V con/tok
6/1
13/0 14/0 15/1 16/1 17/3 18/3
25/(3) 26/3 27/1(2) 28/3 29/2(1) 30/1(2)
37/0 38/2(1) 39/(l) 40/(3) 41/2(1)
49/3 50/0 51/(3) 52/3 53/(3) 54/3
7
8
18
13
15
31/1(2)
43/(3) 44/3 45/2(1) 46/3 47/3 48/3
55/2 56/3 57/3 58/3 59/3 60/3
l/(3)
2/0
3/0 4/3 5/0 Total f
7/(3)
42/(3)
10/3 11/1 12/3
19/1 20/3 21/3 22/3 23/3 24/3
Total m
10
16
18
18
17
Total all
17
24
36
31
32
8/0 9/0
32/(3)
33/2(1) 34/3 35/(3)
36/(3)
con/tok = consultant/token; f = female; m = male. *The number of tokens for the variant pona is given in parentheses.
118
Classificatory Particles in Kilivila
Table 3.2.98. Sa 'nut bunch' (16 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0
49/3 50/0 51/3 52/0 53/0 54/0
38/0 39/0 40/0 41/0 42/0
Total f
6
7/0 8/0 9/0 10/1 11/0 12/0
19/0 20/1 21/0 22/0 23/2 24/0
31/3 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/3
46/0 47/0 48/0
Total m
1
3
3
3
Total all
1
3
3
3
55/0 56/0 57/0 58/0 59/0 60/0
6
con/tok = consultant/token; f = female; m = male.
Table 3.2.99. Si 'small bit' (15 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/1 5/0 6/0
13/2 14/3
25/3 26/1 27/0
37/0 38/0
39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
1
5
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
con/tok
Total f
7/0 8/0 9/0 10/1 11/0 12/3
Total m
4
Total all
5
15/0 16/0 17/0 18/0
28/0 29/0 30/0 4
19/0 20/0 21/0 22/0 23/0 24/0
31/1 32/0 33/0 34/0 35/0 36/0
5
5
1
con/tok = consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
119
Table 3.2.100. Sipu 'sheaf (88 tokens) I con/tok 1/0 2/0 3/0 4/0 5/0 6/0
II
III
con/tok
con/tok
IV con/tok
V con/tok
13/0
25/0
37/0
49/3
14/0
26/1
38/3
50/1
15/1
27/3
39/0
51/3
16/0
28/3
40/3
52/0
41/3
53/3
17/0
9/3
18/0
30/0
42/3
54/3
1
10
12
13
19/1 20/0 21/3 22/0 23/3 24/3
31/3 32/3 33/0 34/3 35/3 36/3
43/3 44/0 45/3 46/3 47/0 48/0
55/3 56/3 57/3 58/3 59/3 60/3
Total m
10
15
9
18
Total all
11
25
21
31
II con/tok
III con/tok
IV con/tok
V con/tok
3/0 4/1 5/0 6/3
13/2 14/5 15/6 16/5 17/6 18/6
25/3 26/3 27/6 28/6 29/6 30/1
37/5 38/6 39/6 40/6 41/5 42/3
49/5 50/6 51/6 52/0 53/6 54/3
4
30
25
31
26
7/0 8/0 9/0 10/3 11/2 12/3
19/6 20/6 21/6 22/3 23/6 24/5
31/5 32/6 33/6 34/6 35/6 36/3
43/6 44/6 45/6 46/3 47/6 48/6
55/6 56/6 57/5 58/4 59/3 60/6
Total m
8
32
32
33
30
Total all
12
62
57
64
56
Total f
7/0 8/0 9/0 10/0 11/0 12/0
con/tok = consultant/token; f = female; m = male.
Table 3.2.101. Sisi 'bough' (251 tokens)
I con/tok 1/0 2/0
Total f
con/tok = consultant/token; f = female; m = male.
120
Classificatory Particles in Kilivila
Table 3.2.102. Siva 'time' (3 tokens)
I con/tok 1/0 2/0 3/0
4/0 5/0 6/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/3 56/0 57/0 58/0 59/0 60/0
Total f
7/0 8/0 9/0 10/0 11/0 12/0 Total m
3
Total all
3
con/tok = consultant/token; f = female; m = male.
Table 3.2.103. Suya 'batch of fish' (90 tokens)
I con/tok 1/0
2/0 3/0 4/0 5/0
6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/0 14/0 15/0 16/0 17/3 18/3
25/0 26/0 27/3 28/0 29/3 30/3
37/0 38/3 39/3 40/3 41/3 42/3
49/3 50/0 51/3 52/3 53/3 54/3
6
9
15
15
43/3 44/3 45/3 46/3 47/3 48/3
55/0 56/3 57/3 58/3 59/0 60/3
19/0 20/0 21/0 22/0 23/0 24/0
Total m Total all
6
con/tok = consultant/token; f = female; m = male.
31/0 32/3 33/3 34/3 35/3 36/3
15
18
12
24
33
27
The System of Classificatory Particles in Kilivila
121
Table 3.2.104. Ta 'basket' (152 tokens)
I con/tok
II con/tok
1/0 2/0 3/0 4/3 5/4 6/3
13/3
III con/tok
IV con/tok
V con/tok
37/2
49/3 50/2 51/3 52/3
17/3 18/3
25/3 26/3 27/1 28/3 29/3 30/3
38/3 39/3 40/3 41/3 42/3
10
16
16
17
17
7/0 8/0 9/0
31/3 32/3 33/3 34/3 35/3 36/2
43/3 44/3 45/2 46/3 47/3 48/3
55/3 56/2
10/3 11/3 12/3
19/3 20/3 21/3 22/3 23/3 24/3
Total m
9
18
17
17
15
Total all
19
34
33
34
32
IV con/tok
V con/tok
37/0
49/0 50/0 51/6
Total f
14/1
15/3 16/3
53/3 54/3
57/2 58/3
59/3 60/2
con/tok = consultant/token; f = female; m = male.
Table 3.2.105. 7am 'sprouting (yams etc)' (70 tokens)
I
con/tok 1/0
II con/tok
2/0
3/0
13/0 14/0 15/0
4/0 5/0 6/0
16/0 17/0 18/0
Total f
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/6 21/1 22/0 23/3 24/0
III con/tok
25/0 26/3 27/0 28/6 29/0 30/0
38/8
39/0 40/3 41/0 42/3
52/0 53/0 54/8
9
14
14
31/0
43/6
32/0 33/0
44/3 45/2 46/0 47/0 48/0
55/0 56/0 57/0 58/0
34/2 35/0 36/6
59/4 60/0
Total m
10
8
11
4
Total all
10
17
25
18
con/tok = consultant/token; f = female; m = male.
122
Classificatory Particles in Kilivila
Table 3.2.106. Tetu 'yams' (1 token)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0
25/0
37/0
49/0
14/0
26/0
38/0
50/0
15/0
27/0
39/0
51/0
16/0
28/0
40/0
52/0
17/0
29/0
41/0
53/0
18/0
30/0
42/0
54/0
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/1 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
con/tok
Total f
10/0 11/0 12/0 Total m
1
Total all
1
con/tok = consultant/token; f = female; m = male.
Table 3.2.107. To/te 'persons of male gender, human beings' (194 tokens)
II con/tok
III con/tok
IV con/tok
V con/tok
4/3 5/3 6/4
13/3 14/3 15/6 16/6 16/3 18/3
25/3 26/3 27/3 28/3 29/3 30/3
37/3 38/3 39/3 40/3 41/3 42/3
49/3 50/3 51/3 52/3 53/3 54/3
19
24
18
18
18
7/3 8/3 9/3 10/3 11/4 12/3
19/6 20/4 21/3 22/3 23/3 24/3
31/3 32/3 33/3 34/3 35/3 36/3
43/3 44/3 45/3 46/3 47/3 48/3
55/5 56/3 57/3 58/3 59/3 60/3
Total m
19
22
18
18
20
Total all
38
46
36
36
38
I con/tok 1/3
2/3 3/3
Total f
con/tok = consultant/token; f = female; m = male.
The System of Classificatory Particles in Kilivila
123
Table 3.2.108. Tuta 'time' (8 tokens)
I
II con/tok
III con/tok
IV con/tok
1/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/3 29/0
37/0 38/0 39/0 40/1 41/0
30/0
42/0
49/0 50/0 51/1 52/0 53/0 54/0
1
1
43/0 44/0 45/2 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
con/tok 2/0 3/0 4/0 5/0 6/0 Total f
28/0
3
7/0 8/0 9/0
19/0 20/0 21/0
31/0 32/0
10/0 11/0
22/0 23/0
34/0
12/0
24/0
33/0 35/1 36/0
V
con/tok
Total m
1
2
Total all
4
3
1
con/tok = consultant/token; f = female; m = male.
Table 3.2.109. Utu 'scrap, parts cut off' (308 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0
13/4
25/7
37/4
2/0 3/0 4/2 5/0 6/2
14/4 15/2
16/2 17/7 18/2
26/9 27/13 28/7 29/6 30/24
38/16 39/5 40/8 41/6 42/6
49/5 50/4 51/6 52/6 53/6 54/3
4
21
66
45
30
7/0 8/0 9/0
31/6 32/6 33/9 34/9 35/9 36/9
43/13 44/9 45/6 46/8 47/9 48/6
55/1
10/3 11/0 12/1
19/1 20/1 21/2 22/1 23/0 24/6
Total m
4
11
48
51
28
Total all
8
32
114
96
58
Total f
con/tok = consultant/token; f = female; m = male.
56/9 57/6 58/2 59/3 60/7
124
Classificatory Particles in Kilivila
Table 3.2.110. Uva 'span' (31 tokens) I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/1 39/0 40/0 41/1 42/0
49/3 50/0 51/0 52/3 53/3 54/1
2
10
2/0 3/0 4/0 5/0 6/0 Total f
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
55/0 56/0 57/0 58/1 59/2 60/3
31/3 32/3 33/0 34/0 35/3 36/0
43/0 44/0 45/3 46/0 47/0 48/1
Total m
9
4
6
Total all
9
6
16
10/0 11/0 12/0
con/tok = consultant/token; f = female; m = male.
Table 3.2. 111. Va 'door' (14 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/1 5/0 6/0
13/0 14/0 15/1 16/6 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/3 42/3
49/0 50/0 51/0 52/0 53/0 54/0
1
7
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
con/tok
Total f
6 31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
Total m Total all
1
7
con/tok — consultant/token; f = female; m = male.
6
55/0 56/0 57/0 58/0 59/0 60/0
The System of Classificatory Particles in Kilivila
125
Table 3.2.112. Vakala 'belt' (2 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/1 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
43/0 44/0 45/1 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
con/tok
5/0
6/0
1
Total f
7/0 8/0 9/0 10/0 11/0 12/0
31/0 32/0 33/0 34/0 35/0 36/0
19/0 20/0 21/0 22/0 23/0 24/0
1
Total m
1
Total all
1
con/tok = consultant/token; f = female; m = male.
Table 3.2.113. Vaya/vayo 'door' (3 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0
25/0
37/0
49/0
14/0
26/0
38/0
50/0
15/0
27/0
39/0
51/0
16/0
28/0
40/0
52/0
17/0
29/0
41/0
53/0
18/0
30/0
42/0
54/0
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/1 35/0 36/0
43/0 44/0 45/2 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
Total f
10/0 11/0 12/0 Total m
1
2
Total all
1
2
con/tok = consultant/token; f = female; m = male.
126
Classificatory Particles in Kilivila
Table 3.2.114. Vili 'untwisted' (76 tokens)
I
con/tok
II con/tok
2/0
1/0
13/0 14/0
3/0
15/0
4/0 5/0 6/0
16/0 17/0 18/0
Total f
7/0 8/0 9/0 10/0 11/0 12/0
19/1 20/0 21/0 22/0 23/0 24/0
IV con/tok
V con/tok 49/1
28/2 29/3 30/3
37/0 38/0 39/5 40/3 41/1 42/5
11
14
13
31/3 32/0 33/3 34/3 35/3 36/3
43/1 44/0 45/3 47/1 48/3
55/0 56/3 57/2 58/3 59/0 60/3
III con/tok
25/0 26/3 27/0
46/3
50/0 51/3 52/3 53/3 54/3
Total m
1
15
11
11
Total all
1
26
25
24
III
IV con/tok
V con/tok
37/0 38/3 39/0
49/0 50/0 51/2 52/0 53/0 54/0
con/tok = consultant/token; f = female; m — male.
Table 3.2.115. Vosi 'song' (18 tokens)
I con/tok
II con/tok
con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/1 27/1 28/0 29/0 30/0
Total f
2
40/3 41/1
42/0 7
31/0 32/0 33/3 34/0 35/1 36/0
43/0 44/0 45/3 46/0 47/0 48/0
Total m
4
3
Total all
6
10
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
con/tok = consultant/token; f = female; m = male.
2
55/0 56/0 57/0 58/0 59/0 60/0
2
The System of Classificatory Particles in Kilivila
127
Table 3.2.116. Wela 'batch of fish' (42 tokens)
I con/tok 1/0 2/0
3/0
4/0 5/0 6/0 Total f
7/0 8/0 9/0 10/0 11/0 12/0
II con/tok
III con/tok
IV con/tok
V con/tok
13/3
25/3 26/0 27/2
49/0 50/1 51/0 52/0
29/0 30/0
37/0 38/3 39/0 40/3 41/3 42/0
3
5
9
19/0 20/0 21/0 22/0 23/0
31/6 32/0 33/0 34/0 35/0 36/3
43/0 44/0 45/3 46/3 47/0 48/3
55/3 56/0 57/0 58/0 59/0 60/3
9
9
6
14
18
7
III
IV con/tok
V con/tok
25/6 26/9 27/1 28/6 29/3
37/0 38/11 39/9 40/4 41/4 42/1
49/13 50/0 51/12 52/0 53/0
14/0 15/0 16/0 17/0 18/0
28/0
24/0
Total m
3
Total all
53/0 54/0
1
con/tok = consultant/token; f = female; m = male.
Table 3.2.117. Ya 'flexible, thin' (273 tokens)
I con/tok 1/0 2/0 3/0 4/6 5/2 6/2
II con/tok 13/7 14/0 15/11
16/7 17/12 18/6
con/tok
30/8
54/6
10
43
33
29
31
7/0 8/0 9/0
19/3 20/15 21/7
31/0 32/3
43/3
10/5 11/4 12/3
22/9 23/5 24/6
34/2 35/3 36/0
45/5 46/0 47/9 48/6
55/4 56/15 57/6 58/3 59/4 60/3
Total m
12
45
10
25
35
Total all
22
88
43
54
66
Total f
con/tok = consultant/token; f = female; m = male.
33/2
44/2
128
Classificatory Particles in Kilivila
Table 3.2.118. Yam1 'day' (19 tokens) con/tok
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/3 27/5 28/0 29/0 30/0
37/0 38/1 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
Total f
8
7/0 8/0 9/0
1
19/0 20/0 21/1 22/0 23/0 24/3
31/3 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/3 46/0 47/0 48/0
Total m
4
3
3
Total all
4
11
4
10/0 11/0 12/0
55/0 56/0 57/0 58/0 59/0 60/0
con/tok = consultant/token; f = female; m — male.
Table 3.2.119. Yom2/yoma/yuma 'hand' (12 tokens)
I
Total f
III
IV
V
con/tok
con/tok
con/tok
con/tok
1/0 2/0 3/0 4/0 5/1 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/2 41/0 42/0
49/0 50/0 51/3 52/0 53/3 54/0
1 7/0 8/0 9/0 10/0 11/0 12/0
2
19/0 20/0 21/2 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
1
2
con/tok = consultant/token; f = female; m = male.
6 55/1 56/0 57/0 58/0 59/0 60/0
1
2
Total m Total all
II
con/tok
2
7
The System of Classificatory Particles in Kilivila
129
Table 3.2.120. Yegila 'name' (7 tokens)
I
con/tok 1/0 2/0 3/0
4/0 5/0 6/0
II con/tok 13/0 14/0 15/0 16/0 17/0 18/0
Total f
IV
con/tok
con/tok
V con/tok
25/0 26/1
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0
III
27/3 28/0 29/0 30/0
54/0
4
7/0 8/0 9/0 10/0
11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0
32/0 33/0 34/0 35/2 36/1
Total m
3
Total all
7
60/0
con/tok = consultant/token; f = female; m = male.
Table 3.2.121. Yulai/yule 'bundle of 4 things' (9 tokens)
I
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
7/0 8/0 9/0
19/0
31/3 32/0 33/0
10/0 11/0 12/0
22/0 23/0 24/0
35/3
43/0 44/0 45/0 46/0 47/0
55/3
20/0 21/0
36/0
48/0
con/tok
27/0 28/0 29/0 30/0
Total f
34/0
56/0 57/0 58/0 59/0 60/0
Total m
6
3
Total all
6
3
con/tok — consultant/token; f = female; m = male.
130
Classificatory Particles in Kilivila
Table 3.2.122. Yuva 'shoal' (4 tokens)
I
II
III
IV
V
con/tok
con/tok
con/tok
con/tok
con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/3
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/0 54/0
43/0 44/0 45/0 46/0 47/0 48/0
55/1 56/0 57/0 58/0 59/0 60/0
Total f
3
7/0 8/0 9/0 10/0 11/0 12/0
19/0 20/0 21/0 22/0 23/0 24/0
31/0 32/0 33/0 34/0 35/0 36/0
1
Total m
1
3
Total all con/tok = consultant/token; f = female; m = male.
Table 3.2.123. 0 'basketful of yams' (2 tokens)
I con/tok
II con/tok
III con/tok
IV con/tok
V con/tok
1/0 2/0 3/0 4/0 5/0 6/0
13/0 14/0 15/0 16/0 17/0 18/0
25/0 26/0 27/0 28/0 29/0 30/0
37/0 38/0 39/0 40/0 41/0 42/0
49/0 50/0 51/0 52/0 53/1 54/0
7/0 8/0 9/0
19/0 20/0 21/0 22/0 23/0 24/0
31/1 32/0 33/0 34/0 35/0 36/0
43/0 44/0 45/0 46/0 47/0 48/0
55/0 56/0 57/0 58/0 59/0 60/0
1
Total f
10/0 11/0 12/0 Total m
1
Total all
1
con/tok = consultant/token; f = female; m = male.
1
The System of Classificatory Particles in Kilivila
131
At this stage of the data presentation, I have assigned—provisionally—to all classifiers listed in the table headings the status "CP type." In the analyses that follow the data presentation, I will define the concept "CP type" and decide which of these classifiers are CP types proper and which are variants of CP types. Table 3.2.124 presents the number of the provisionally defined types produced by each of the 60 informants; it also presents the average resulting from the subtraction of the average for the types produced by the female informants from the average for the types produced by the male informants (see also section 3.2.1). What strikes even the casual reader of Tables 3.2.37 to 3.2.123 is the wide spectrum of tokens produced, ranging from 1 to 2,877. Table 3.2.125 lists the 87 CP types elicited, ordering them according to the number of tokens produced for each CP type. The table also presents percentages based on the calculus that (with the exception of the CP 0) 180 tokens of each CP type produced by 60 consultants with three word classes equals 100% of elicitation success. With the 87 types, the consultants produced 10,583 tokens altogether in the elicitation test.
Table 3.2.124. Consultants and number of CP types produced
I con/ types
II con/ types
III con/ types
IV con/ types
V con/ types
1/8 2/2
13/22 14/16
3/2
15/33 16/21 17/32 18/24
37/18 38/43 39/33 40/44 41/47 42/38
49/41 50/32 51/41
4/18 5/12 6/16
25/31 26/42 27/48 28/40 29/30 30/35
9.7
24.7
37.7
37.2
35.7
7/6 8/5 9/2 12/20
19/34 20/29 21/35 22/29 23/21 24/31
31/47 32/35 33/37 34/43 35/46 36/35
43/35 44/32 45/56 46/37 47/31 48/34
55/44 56/37 57/38 58/41 59/37 60/42
Average m
13.5
29.8
40.5
37.5
39.8
Average all
11.6
27.3
39.1
37.3
37.8
3.8
5.1
2.8
0.3
4.1
Average f
10/29
11/19
Average m — Average f
con/types = consultant/CP types; f = female; m = male
52/27
53/36 54/37
132
Classificatory Particles in Kilivila
Table 3.2.125. CP types and tokens—all consultants (elicitation test) CP Type kwe 'thing' ke 'wood' kada 'road' utu 'scrap' na2 'animals' ya 'flexible' sisi 'bough' pila 'part' kweya 'limb' to/te 'male' na1 'female' kasa 'row' kumla 'earth oven' kova 'fire' budalboda 'group' ta 'basket' gudi 'child' iga 'name' pwaninalpona 'hole' kai 'stone blade' ka'i 'tooth' duli 'cluster' giwi 'cut' notulnutu 'kneaded, dot' kwoila 'clay pot' kubila 'land plot' deli 'group' suya 'batch of fish' gum 'bit' sipu 'sheaf kabisi 'compartment' luba 'bundle' vili 'untwisted' liku 'compartment' tarn 'sprouting' bubulbobu 'cut across' hwa 'tree' gula 'heap' po'ula 'grove'
Tokens
Percent (180 tokens + = 100%)
2,877 1,919 341 308 279 273 251 249 215 194 184 165 162 159 156 152 142 142 140 136 127 108 106 103 96 94 93 90 88 88 78 77 76 71 70 68 61 56 53
1,598 1,066 189 171 155 152 139 138 119 108 102 92 90 88 87 84 79 79 78 76 71 60 59 57 53 52 52 50 49 49 43 43 42 39 39 38 34 31 29
The System of Classificatory Particles in Kilivila
133
Table 3.2.125. (continued) CP Type
Tokens
kapwalkapo 'parcel' wela 'batch of fish'
46
uva 'span'
mmwa 'bundle' nunu 'corner/garden'
31 29 26 26 23 22
kabulo 'sector'
21
yam 'day' vosi 'song' luva 'wooden dishes'
19 18 18 16
bubwa'cut off kudu 'tooth' kwoya 'mountain'
sa 'nut bunch' kaduyo 'entrance' si 'small bit' va 'door' bogi 'night' gili 'row' megwa 'magic' yama/yuma 'hand' kudu 'band of fibers' yulai 'bundle of four things' tuta 'time' kila 'hands of bananas' lilollola 'walk' peta 'basket'
42
15
Percent (180 tokens + = 100%) 26 23 17 16 14 14 13 12 12 11
10 10 9 8
15 14
8
12 12 12 12 9 9 8
7
7 7
8 7 7 7 5
5 4 4
7
4 4
7
4
yegila 'name' bwalita 'sea' 0 'basketful of yams' duya 'door'
6 2
3
4
2
yuva 'shoal'
4
bukwa 'fruit cluster' guba 'bundle/taro'
3 3 3
2 2 2
kabulo 'point' kala 'day' kavi 'tool' meila 'song/part' nina 'song/part' siva 'time'
3
3 3 3 3
2 2 2 2 2
3
2
134
Table 3.2.125. CP Type
Classificatory Particles in Kilivila
(continued) Tokens
vaya/vayo 'door'
3
beku 'stone blade'
2
vakala 'belt'
2 1
(*)boma '???' doba 'skirt'
1
(*)iki '???'
1
(*)kava '???' nigwa 'hole'
1 1
tetu 'yams'
1
Percent (180 tokens + = 100%) 2 1 1
0.6 0.6 0.6 0.6 0.6 0.6
The CP 0 is only produced with numerals to count baskets of yams; thus, in connection with this CP, 60 tokens produced by 60 consultants during the test would total 100% of elicitation success.
3.2.2.2 CP Types/Tokens Produced Versus CP Types Expected During the Elicitation Test The tables presented so far show that a number of CPs were produced by the consultants as a response to stimuli that should have elicited other CPs. Tables 3.2.126 to 3.2.212 present all the CP types produced by the consultants of the five age groups during the elicitation test allocating the tokens of each CP type produced to the CP type or types actually expected. When necessary, table footnotes provide the following information: first, which other CP(s) were produced when the CP under consideration was expected and second brief comments on the CP produced and/or on those reactions of the consultants that were idiosyncratic, dubious, or difficult to understand. If we compare the data in these tables with the expectations that were the basis for compiling the questionnaire, we note the following: 1. Four types of expected CPs were not realized by the consultants during the elicitation test: a. Lila 'bough, branch, leaf. The CP ke 'wood, long objects' (I-V; 125 tokens) was produced instead. b. Oyla 'string/fish.' The CPs ke 'wood, inanimates' (I-III, V; 34 tokens), ka'i* 'stone blade' (II; 3 tokens), kwe 'thing' (II-IV; 14 tokens), na2 'animals' (II-V; 53 tokens), pila* 'part' (II, V; 2 tokens), and wela 'batch of fish' (II-V; 30 tokens) were produced instead. c. Siwa 'sea portions.' The CPs bwalita 'sea' (III, IV; 6 tokens), kwe 'thing' (I-V; 137 tokens), and utu* 'scrap' (III; 3 tokens) were produced instead. d. Vilo 'place, village.' The CPs kada* 'road' (II; 2 tokens) and kwe 'thing' (I-V; 153 tokens) were produced instead. 2. A number of CPs that were produced were not expected; Table 3.2.213 lists these CPs. Moreover, it should be noted that consultants of age
The System of Classificatory Particles in Kilivila
135
group IV produced the CP luva in its meaning 'tied bundle' instead of the expected CPs bukwa (2 tokens), mmwa (3 tokens), and sa (3 tokens). The production of the CP luva in its meaning 'tied bundle' was not expected, either. 3. The data documenting dubious cases of CP production are presented in Table 3.2.214. These cases of dubious (including not acceptable) CP production are distributed across the sexes as shown in Table 3.2.215. Table 3.2.126. Beku 'stone blade' Tokens of CP beku produced
by age group
1
CP expected
II
kai 'stone blade'
III
IV
1
1
V
Tokens total 2
Beku is a repeater; the production of this CP was not expected.
Table 3.2.127. Bog! 'night' Tokens of CP bogi produced by age group CP expected
I
bogi 'night'
II
III
IV
2
8
2
V
Tokens total
12
1. CP produced, bogi expected: kwe (I-V; 138 tokens). 2. Bogi is a repeater.
Table 3.2.128. Boma '???' Tokens of CP (*) boma produced by age group CP expected
I
II
utu 'scrap'
III
IV
V
1
Tokens total
1
There is no CP boma in Kilivila.
Table 3.2.129. Bubu/bobu 'block cut off/cut across' Tokens of CP bubu/bobu produced by age group CP expected bubitlbobu 'cut across'
I
II
III
IV
V
Tokens total
7
23
13
25
68
CPs produced, bubulbobu expected: bubwa (TV; 2 tokens), buda* (II, III; 4 tokens), bwa (IV; 3 tokens), gula* (II,III, 3 tokens), ke (I, III V; 18 tokens), kudul* (V; 1 token), kwe (II V; 28 tokens), nal* (I; 1 token), ya (I, II; 4 tokens).
136
Classificatory Particles in Kilivila
Table 3.2.130. Bubwa 'parts cut off' Tokens of CP bubwa produced by age group CP expected
II
I
bubwa 'cut off utu 'scrap'
1
III
IV
V
6
5 1
5
1
kabulol 'point'
5
1
vili 'untwisted' Total
1
8
16 8
(1*)
2 2
bubu 'cut across'
Tokens total
10
2 (2*)
10
1. CPs produced, bubwa expected: bwa (I-V; 16 tokens), ke (I-V; 130 tokens), kwe (IV; 1 token), utu (IV; 2 tokens). 2.1 can justify the production of the CP bubwa as a response to the stimulus presented to elicit the CP kabulol, assuming that the consultant thought of a 'protruding piece of wood that is cut off.
Table 3.2.131. Buda/boda/budu 'group, team' Tokens of CP buda/boda/budu produced by age group CP expected
V
Tokens total
12
17
63
18
27
86
II
III
IV
3
7
24
5
11
25
bubu 'cut across'
1
3
gum 'bit'
1
buda/boda/budu
I 'group'
yuva 'shoal'
(4*) (1*)
gula 'heap, group'
1
mmwa 'conical bundle'
1
Total
8
20
52
32
1 1 44
1. CP produced, buda expected: kwe (I-V; 85 tokens). 2. See table 3.2.166 below; see the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.132. Bukwa 'fruit cluster' Tokens of CP bukwa produced by age group CP expected
I
bukwa 'fruit cluster'
3
II
III
IV
V
Tokens total 3
CPs produced, bukwa expected: ke (I V; 122 tokens), kwe (II, III; 4 tokens), luva* (IV; 2 tokens), po'ula (V; 1 token), sa (IV; 3 tokens), ya (II, IV, V; 14 tokens).
The System of Classificatory Particles in Kilivila
137
Table 3.2.133. Bwa 'tree, wood' Tokens of CP bwa produced by age group CP expected
I
II
III
IV
V
Tokens total
utu 'scrap' bubwa 'cut off
3 1
11 2
1 10
1 3
4
20 16
bwa 'tree, wood' kabisi 'compartment' kabulol 'point' k e 'wood'
2
2
1
5
3
1
2 1
1
vili 'untwisted'
1
si 'small bit' bubu 'cut across'
6
Total
4 3 2
1 1
3 10
26
12
2
6 3
7
6
1. CP produced, bwa expected: ke (I-V; 155 tokens). 2. In some cases, this CP was produced because some consultants referred only to the noun and not to its modification within the noun phrases presented as stimuli.
Table 3.2.134. Bwalita 'sea' Tokens of CP bwalita produced by age group CP expected
I
II
siwa 'sea portions'
III
IV
2
4
V
Tokens total 6
The CP bwalita is a repeater; the production of this CP was not expected.
Table 3.2.135. Deli 'group on the move' Tokens of CP deli produced by age group CP expected deli 'group'
I
II
III
IV
V
Tokens total
4
30
23
36
93
1. CPs produced, deli expected: kasa (III; 2 tokens), kwe (I IV; 37 tokens), nal (I; 1 token), to/te (I, II; 11 tokens). 2. See also Tables 3.2.156 (kasa) and 3.2.166 (kwe).
138
Classificatory Particles in Kilivila
Table 3.2.136. Doba '(grass) skirt' Tokens of CP doba produced by age group CP expected
I
II
III
kudu 'band of fibres'
IV
V
1
Tokens total 1
The CP doba is a repeater; the production of this CP was not expected.
Table 3.2.137. Duli 'cluster, bundle' Tokens of CP duli produced by age group CP expected
I
II
III
IV
V
Tokens total
duli 'cluster'
3
15
33
27
30
108
V
Tokens total
1
4
CPs produced, duli expected: ke (II; 6 tokens), kwe (I, II, IV; 29 tokens).
Table 3.2.138. Duya 'door, entrance' Tokens of CP duya produced by age group CP expected
I
II
duya 'door, entrance'
III
IV
3
CPs produced, duya expected: kada (I-V; 70 tokens), ke (II; 5 tokens), kwe (II, V; 21 tokens), ya (II, IV, V; 28 tokens), nal* (III; 3 tokens), va (II, IV; 10 tokens), vayalvayo (III, IV; 2 tokens), kaduyo (HI, V; 8 tokens).
Table 3.2.139. Gili 'row' Tokens of CP gili produced by age group CP expected gili 'row'
I
II
III
IV
V
Tokens total
3
1
8
12
CPs produced, gili expected: kasa (11, III; 3 tokens), kwe (II V; 29 tokens), utu* (II; 1 token), vili* (IV; 3 tokens), ya (I V; 101 tokens), vakala (III, IV; 2 tokens).
The System of Classificatory Particles in Kilivila
139
Table 3.2.140. G/wTcut' Tokens of CP giwi produced by age group CP expected
I
giwi'cut'
II
III
IV
V
Tokens total
17
33
31
25
106
CPs produced, giwi expected: kasa* (I; 3 tokens), ke (I-V; 16 tokens), kwe (II, IV, V; 11 tokens), luva* (II; 3 tokens), na2* (II; 1 token), to* (I; 1 token).
Table 3.2.141. Guba 'bundle of taro' Tokens of CP guba produced by age group CP expected
I
II
III
IV
mmwa 'conical bundle'
V
Tokens total
3
3
The production of the CP guba was not expected.
Table 3.2.142. Gudi 'child' Tokens of CP gudi produced by age group CP expected
I
II
III
IV
V
Tokens total
gudi 'child'
7
30
36
33
36
142
Table 3.2.143. Gula 'heap, group' Tokens of CP gula produced by age group
II
III
IV
V
Tokens total
gula 'heap, group'
6
11
20
13
50
bubu 'cut across'
2
1
(3*)
3
3
CP expected
I
po'ula 'grove' Total
8
15
20
13
1. CPs produced, gula expected: buda (IV; 1 token), kwe (1-V; 82 tokens), tam (II-V; 28 tokens). 2. Sec the comments on dubious (*) CP usage (see section 3.2.1).
140
Classificatory Particles in Kilivila
Table 3.2.144. Gum 'bit, small piece' Tokens of CP gum produced by age group CP expected
I
gum 'bit, small piece'
II
III
IV
V
Tokens total
9
30
28
21
88
CPs produced, gum expected: buda* (II; 1 token), kasa* (II; 3 tokens), ke (IV, V; 6 tokens), kwe (I, II, IV, V; 22 tokens), na2* (I; 3 tokens), utu (II; 1 token), ya (II; 6 tokens).
Table 3.2.145. Iga 'name' Tokens of CP iga produced by age group CP expected iga 'name'
I
II
III
IV
V
Tokens total
10
34
29
33
36
142
CPs produced, iga expected: kasa* (I; 1 token), kwe (I, IV; 6 tokens), yegila (III; 7 tokens).
Table 3.2.146. lki'???' Tokens of CP "Iki produced by age group
CP expected
I
nina 'part of a song'
II
III
IV
V
Tokens total
l
1
There is no CP iki in Kilivila; it may be that the consultant wanted to produce the CP igi 'wind': if we assume that the consultant wanted to refer to wind or rain magic as a reaction to the stimulus megwa 'magic', we may understand this reaction, although it is highly idiosyncratic (see also G. Senft 1985c).
Table 3.2.147. Kabisi 'compartment of a foodhouse' Tokens of CP kabisi produced by age group CP expected kabisi 'compartment'
I
II
III
IV
V
Tokens total
10
23
23
22
78
CPs produced, kabisi expected: hwa (I, II; 4 tokens), ke (I-V; 53 tokens), kwe (I -IV; 10 tokens), nunu* (II: 1 token), pila (I; 1 token), utu (II, III; 4 tokens).
141
The System of Classificatory Particles in Kilivila Table 3.2.148. Kabulol 'protuberances, point' Tokens of CP kabulol produced by age group CP expected
I
II
III
IV
nunu 'corner of a garden'
V
Tokens total
3
(*3)
1. CPs produced, kabulol expected: bubwa* (III; 1 token), bwa (II, III; 3 tokens), ke (I-V; 107 tokens), utu (IIV; 38 tokens). 2. The CP kabulo was produced as a response to the stimulus presented to elicit the expected CP nunu; this can be explained by assuming that the speaker either (1) mixed up the CPs kabulol and kabululkabulo2, extending the meaning of kabulo2 from 'village sector' to 'sector in general' and then responding to the stimulus presented to elicit the CP nunu with the CP kabulol in its extended meaning, or (2) wanted to use the meaning 'point' of this CP to refer to a 'corner of a graden'.
Table 3.2.149. Kabulu/kabulo2 Village sector' Tokens of CP kabulu/kabulo2 produced by age group CP expected
I
kabulu 'village sector'
II
III
IV
V
Tokens total
2
10
2
7
21
CPs produced, kabulu expected: kada' * (III; 3 tokens), ke (I-V; 92 tokens), kwe (II-V; 40 tokens), utu (II; 1 token).
Table 3.2. 150. Kada/keda 'road, track' Tokens of CP kada/keda produced by age group CP expected
I
II
III
IV
V
Tokens total
kada 'road, track'
8
36
36
36
36
duya 'door, entrance'
5
20
21
14
10
152 70
kaduyo 'entrance, door'
4
22
33
28
26
113
lilo 'walk, journey'
1
vilo 'place, village'
2
(2*) 3
kabulu 'village sector' Total
(1*)
17
81
93
See the comments on dubious (*) CP usage (see section 3.2.1).
(3*) 78
72
142
Classificatory Particles in Kilivila
Table 3.2.151. Kaduyo/kaduya 'entrance, door' Tokens of CP kaduyo produced by age group CP expected
I
II
III
IV
V
Tokens total
kaduyo 'door, entrance'
1
1
3
1
1
7
3
8
5
duya 'door, entrance' Total
1
1
8
1
4
CPs produced, kaduyo expected: kada (I-V; 113 tokens), ke (II; 4 tokens), kwe (II, V; 4 tokens), ya (II, IV, V; 17 tokens), va (I, II; 4 tokens), vaya/vayo (IV; 1 token).
Table 3.2.152. Kai 'stone blade' Tokens of CP kai produced by age group CP expected
I
II
III
IV
V
Tokens total
kai 'stone blade'
3
27
35
35
36
136
CPs produced, kai expected: kwe (II; 2 tokens), nal* (II; 1 token), sisi* (II; 2 tokens), beku (III, IV; 2 tokens).
Table 3.2.153. Ka'/'tooth' Tokens of CP ka'i produced by age group CP expected kudu2 'tooth'
I
II
III
IV
V
Tokens total
13
27
28
26
30
124
Total
3*
3
oyla 'string'
13
30
28
26
30
The production of the CPfaj'i was not expected, see the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.154. Kala 'day' Tokens of CP kala produced by age group CP expected kala 'day'
I
II
III
IV
V
Tokens total
3
3
CPs produced, kala expected: kwe (J-V; 144 tokens), lulu (IV; 1 token), yam (III, IV; 5 tokens).
The System of Classificatory Particles in Kilivila
143
Table 3.2.155. Kapwa "bundle/parcel wrapped up' Tokens of CP kapwa produced by age group CP expected
I
II
III
IV
V
Tokens total
kapwa 'parcel'
5
18
6
6
11
46
CPs produced, kapwa expected: kwe (I-V; 30 tokens), luba (I-V; 77 tokens).
Table 3.2.156. Kasa 'row, line' Tokens of CP kasa produced by age group CP expected
I
II
III
IV
V
Tokens total
kasa 'row, line'
10
24
36
33
36
139
3
4
3
10
meila 'song/part'
1
gili 'row' iga 'name'
2
3
1
(1*)
gtwi 'cut'
3
(3*)
gum 'bit'
3
(3*)
yuva 'shoal'
1
(1*)
nina 'song/part'
3
3
deli 'group'
2
2
Total
11
32
46
37
39
1. CPs produced, kasa expected: ke (I; 2 tokens), kwe (II, IV; 9 tokens). 2. See the comments on dubious (*) CP usage (see section 3.2.1); the CP kasa was produced as a response to the stimuli presented to elicit the expected CPs meila, nina, and deli; assuming that the consultants thought of 'a line as part of a song' and of 'a group on the move, walking one behind the other in a row', we can justify the production of this CP.
Table 3.2.157. Kava'???' Tokens of CP *kava produced by age group CP expected kwoila 'clay pot' There is no CP kava in Kilivila.
I
II
III
IV
V
Tokens total
1
1
144
Classificatory Particles in Kilivila
Table 3.2.158. /Cow'tool' Tokens of CP kavi produced by age group CP expected
I
II
III
IV
kavi 'tool'
V
Tokens total
3
3
CPs produced, kavi expected: ke (II; 7 tokens), kwe (II; 6 tokens), pila* (I; 1 token), utu* (II; 1 token).
Table 3.2.159. Kel-4 Wood (1), long objects (2), inanimate (3), fire' (4) Tokens of CP ke produced by age group CP expected
I
II
III
ke 'wood' bubwa 'cut off
23 12
35 36
bukwa 'fruit cluster' bwa 'tree'
9 14
IV
V
Tokens total
36
36 31
36 31
166
20
24
33
25
34
36
36
31 35
130 122 155
giwi 'cut'
4
2
3
1
6
16
kabulo 'point' kabulu 'sector'
8
32
12
26
29
107
15 9
21
12
26
33
36
29
18 35
92 142
8 10
36
36
36
36
36
36 33
152
34
15
33
27
28
18
8 5
22
33
25
18
3
4
26 6
7 3
34
27
25
1
3
3
6
12
9
12
3
2
17
12
128
1
9 3 13
11
2
kila 'hands/bananas' lila 'bough' luva 'wooden dishes' mmwa 'conical bundle' sa 'nut bunch' utu 'scrap' uva 'span' ya 'flexible, thin' kabisi 'compartment' liku 'compartment' vili 'untwisted' si 'small bit' bubu 'cut across' oyla 'string' sisi 'bough' meila 'song/part' gum 'bit' duya 'door' kova Tire' kubila 'land plot' kwoila 'clay pot'
6 10
3
9
2
2
6 3 3 1 1
2
4
1
17 8 14 10 8 6 6 2 4 6 3
149 121 114 36 110 18 53 32 49 28 18 34 8 12 6 5 1 1 5
145
The System of Classificatory Particles in Kilivila Table 3.2.159. (continued) Tokens of CP ke produced by age group CP expected
I
II
pila 'part, piece'
2
yulai 'bundle of four things'
1
III
IV
V
Tokens total
3
5
1
2
kweya 'limb'
5
5
kudu 'tooth'
2
2
kavi 'tool'
7
7
kaduyo 'entrance'
4
4
6
6
duli 'cluster' yuva 'shoal'
1
1
sipu 'sheaf
2
2
Wo 'walk'
1
1
kasa 'row'
2
2
Total
204
499
380
412
424
1. CP produced, ke expected: bwa (I, II; 2 tokens). 2. The CP yuva, which was replaced by ke, normally implies animation of the objects referred to.
Table 3.2.160. Kila 'clusters/hands of bananas' Tokens of CP kila produced by age group CP expected
I
kila 'hands o f bananas'
2
II
III
IV
V
Tokens total
4
1
7
CPs produced, kila expected: ke (I-V; 142 tokens), kwe (I, II, IV; 8 tokens).
Table 3.2.161. Kova 'fire' Tokens of CP kova produced by age group CP expected kova 'fire' yuma 'hand' Total
I
II
III
IV
V
Tokens total
16
36
36
35
35
158
1 17
(1*) 36
36
1. CPs produced, kova expected: ke (IV; 1 token), kwe (V; 1 token). 2. See the comments on dubious (*) CP usage (see section 3.2.1).
35
35
146
Classificatory Particles in Kilivila
Table 3.2.162. Kabila 'large land plot1 Tokens of CP kubila produced by age group CP expected
I
kubila 'land plot'
II
III
IV
V
Tokens total
12
26
29
27
94
CPs produced, kubila expected: ke (IV; 1 token), kwe (I-V; 53 tokens), pila (I; 1 token), utu* (III; 2 tokens).
Table 3.2.163. Kudul 'band of fibers' Tokens of CP kudul produced by age group CP expected
II
I
III
IV
V
3
Tokens total
3
6
bubu 'cut across'
1
yuva 'shoal'
2
(1*) (2*)
IV
V
Tokens total
9
6
kudu 'fibers'
1. CPs produced, kudul expected: kwe (I-V; 145 tokens), doba (IV; 1 token). 2. See the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.164. Kadu2 'tooth' Tokens of CP kudu2 produced by age group CP expected
I
II
kudu 'tooth'
3
III 8
2 6
CPs produced, kudu2 expected: ke (II; 2 tokens), kwe (II, IV; 2 tokens), ka'i (I-V; 127 tokens).
Table 3.2.165. /Cum/a'earth oven' Tokens of CP kumla produced by age group CP expected
I
II
III
IV
V
Tokens total
kumla 'earth oven'
19
36
36
36
35
162
CP produced, kumla expected: kwe (V; 1 token).
Ghe System of Classificatory Particles in Kilivila
147
Table 3.2.166. Kwe 'thing, unmarked form for inanimates' Tokens of CP kwe produced by age group III
IV
V
Tokens total
36
36
36
25
34
36 36
164
28
13 14
24
16
7
17
85 82
35 7 20
30
33
33
144
9 8
5 6
4
30 53
34
36 21
CP expected
I
kwe 'thing' bogi 'night' buda 'group' gula 'heap'
20 15 6
26
17
27
kala 'day' kapwa 'parcel' kubila 'land plot' kudu 'fibers' kwoila 'clay pot' kwoya 'mountain' lilo 'walk' nina 'song/part' nunu 'corner/garden'
13 5 10 10 2
po'ula 'grove' sipu 'sheaf siva 'time' siwa 'sea portions' vilo 'place' yam 'day' yulai 'bundle of four things' yuva 'shoal' 0 'basketful of yams' bubu 'cut across' duya 'door' gili 'row' kabulu 'sector' gum 'bit' liku 'compartment' notu 'kneaded, dot' pwanina 'hole' sa 'nut bunch' giwi 'cut' vili 'untwisted' kaduyo 'entrance' mmwa 'conical bundle' kova fire'
II
14 12
6 31
138
32
9 33
145
25
10 29
10 33
49 132
29 25
27 10
35 10
28 9
131
7 10 12
32 28
28 12
27
27 14
124
5
17 36
10 29
49 134
28 33
28 36
28 32 36
5 31
26
27
34
36 36 36
137 153 141
35
30 7
36 18
33
148 52
29 3 3
23 12
10 13 12 18 14 4 15
20 27 12 4 4
2 14 1 3 2
10 12 22 15 12 9 5 2 1 1
9 11
27 12
1 12 8 5
12 12 5 6 4 1
3 23 1 13 4 11 3 7 5 3 3 2 1 3 3 1
61 93
117 28 21 29 40 22 55 33 23 20 11 4 4 4 1
148
Classificatory Particles in Kilivila
Table 3.2.166. (continued) Tokens of CP kwe produced by age group CP expected
I
II
III
IV
V
Tokens total
kumla 'earth oven'
1
1
pila 'part, piece'
3
3
deli 'group'
1
kabisi 'compartment'
19
4 3
13 1
37 10
1
5
11
9
9
29
kila 'hands/bananas'
4
1
meila 'song/part'
2
5
3 2
8 9 14
duli 'cluster'
oyla 'string'
5
iga 'name'
3
6 1
kudu 'tooth'
6
3 3 1
3
kasa 'row'
6
9
bubwa 'cut off
1
2 1
to 'basket'
1
1
yuma 'hand'
11
33
12
1
3
bukwa 'fruit cluster' uva 'span'
56 4
kavi 'tool'
2 6
kweya 'limb'
9
8 2 6 9
luva 'wooden dishes'
2
2
si 'small bit'
3
3
6
2
kai 'stone blade'
Total
305
804
553
643
572
The CPs buda, deli, and yuva, which were replaced by the CP kwe, normally imply animation of the subjects referred to.
Table 3.2.167. Kweya/kwaya/keya
'limb'
Tokens of CP kweya produced by age group CP expected
I
II
kweya 'limb'
3
20
yuma 'hand' Total
3
20
III
IV
V
Tokens total
131 84
36
36
36
24
33
27
60
69
63
CPs produced1, kweya expected: ke (II; 5 tokens), kwe (II; 9 tokens), nal* (I, II; 4 tokens),
The System of Classificatory Particles in Kilivila
149
Table 3.2.168. Kwoila/kwela "clay pot' Tokens of CP kwoila produced by age group CP expected
I
II
III
IV
V
Tokens total
kwoila 'clay pot'
5
26
14
26
25
96
CPs produced, kwoila expected: ke (II, III; 5 tokens), kwe (I-V; 49 tokens), kava* (V; 1 token).
Table 3.2.169. Kwoya/koya "mountain, hill' Tokens of CP kwoya produced by age group CP expected
I
II
III
IV
V
Tokens total
kwoya 'mountain'
1
4
11
7
3
26
CPs produced, kwoya expected: kwe (I-V; 132 tokens).
Table 3.2.170. Liku 'compartment of a foodhouse' Tokens of CP liku produced by age group CP expected
I
liku 'compartment'
II
III
IV
V
Tokens total
11
22
19
19
71
CPs produced, liku expected: ke (II-V; 32 tokens), kwe (I-III, V; 55 tokens).
Table 3.2.171. Lilo/lola 'walk, journey' Tokens of CP lilo/lola produced by age group CP expected
I
lilo 'walk, journey'
1
II
III
IV
V
Tokens total
3
1
2
7
CPs produced, lilo expected: kada* (I; 1 token), ke (I; 1 token), kwe (I-V; 131 tokens), ya (II, III, V; 16 tokens). Table 3.2.172. Luba 'bundle' Tokens of CP luba produced by age group CP expected
I
II
III
IV
V
Tokens total
kapwa 'bundle, parcel'
1
9
21
25
21
77
This CP was not expected; the CP luba is a repeater.
150
Classificatory Particles in Kilivila
Table 3.2.173. Luva1 + 2 'wooden dishes, tied bundle' Tokens of CP luva produced by age group
II
CP expected
I
luval 'wooden dishes'
1
III
IV
V 3
3
giwi 'cut'
Tokens total
4
bukwa 'fruit cluster'
2
(3*) (2*)
mmwa 'conical bundle'
3
(3*)
sa 'nut-bunch'
3
(3*)
3
utu 'scrap' Total
1
3
8
(3*)
6
1. CPs produced, luva expected: ke (I-V; 149 tokens), kwe (II; 2 tokens). 2. See the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.174. Megwa 'magic' Tokens of CP megwa produced by age group CP expected
I
nina 'part of a song'
II
III
IV
1
5
6
V
Tokens total
12
This CP was not expected; the CP megwa is a repeater.
Table 3.2.175. Meila 'part of a song' Tokens of CP meila produced by age group CP expected
I
II
III
IV
meila 'part of a song'
V
Tokens total
3
3
CPs produced, meila expected: kasa (III-V; 10 tokens), ke (II, V; 12 tokens), kwe (I, II, IV; 9 tokens), pila (I-V; 91 tokens), utu (III; 4 tokens), vosi (III-V; 18 tokens).
Table 3.2.176. Mmwa 'conical bundle' Tokens of CP mmwa produced by age group CP expected mmwa 'conical bundle'
I
II
Ml
IV
V
Tokens total
2
8
1
12
23
CPs produced, mmwa expected: buda (IV; 1 token), ke (I-V; 121 tokens), kwe (II, V; 4 tokens), luva* (IV; 3 tokens), utu* (III, IV; 4 tokens), guba (V; 3 tokens).
151
The System of Classificatory Particles in Kilivila Table 3.2.177. Nal 'persons of female gender' Tokens of CP nal produced by age group CP expected nal 'female persons'
I
II
III
IV
V
Tokens total
36
36
36
173
30
35
kweya 'limb'
1
3
bubu 'cut across'
1
(4*) (1*)
deli 'group'
1
1
uva 'span'
1
(1*)
kai 'stone blade'
1
duya 'door' Total
(1*) 3
34
39
39
(3*) 36
36
See the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.178. Na2 'animals' Tokens of CP na2 produced by age group CP expected
I
II
na2 'animals'
18
36
36
3
12
27
6
3
oyla 'string (of fish)' wela 'batch of fish' gum 'bit, small piece'
18
IV
V
Tokens total
36
36
162
12
26
53
6
60
3
giwi 'cut' Total
III
(3*) 1
39
67
(1*) 54
51
68
See the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.179. Nigwa 'hole, nest' Tokens of CP nigwa produced by age group CP expected pwaninalpona 'hole' This CP was not expected.
I
II
III
IV
V
Tokens total
1
1
152
Classificatory Particles in Kilivila
Table 3.2.180. Nina 'part of a song' Tokens of CP nina produced by age group CP expected
I
II
III
IV
nina 'part of a song'
V
Tokens total
3
3
CPs produced, nina expected: *iki (II; 1 token), kasa (III; 3 tokens), kwe (I-V; 61 tokens), pila (II-V; 67 tokens), utu (III; 3 tokens), megwa (II-IV; 12 tokens).
Table 3.2.181. Nutu/notu 'kneaded, dot, drop' Tokens of CP nutu/notu produced by age group CP expected
I
II
III
IV
V
Tokens total
notu 'kneaded, dot'
3
12
33
24
31
103
Tokens total
CP produced, nutu/notu expected: kwe (I, II, IV, V; 33 tokens).
Table 3.2.182. Nunu 'corner of a garden' Tokens of CP nunu produced by age group CP expected
I
II
nunu 'corner/garden' kabisi 'compartment'
1
Total
1
III
IV
V
6
9
6
6
9
6
21 (1*)
1. CPs produced, nunu expected: kabulol* (V; 3 tokens), kwe (I-V; 124 tokens), utu (III; 2 tokens). 2. See the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.183. Peta 'basketTokens of CP peta produced by age group CP expected
I
II
t a 'basket' This CP was not expected; the CP peta is a repeater.
III
IV
V
Tokens total
3
1
3
7
153
The System of Classificatory Particles in Kilivila Table 3.2.184. Pila 'part, piece' Tokens of CP pila produced by age group CP expected
I
II
HI
IV
V
Tokens total
pila 'part, piece' meila 'part/song'
3
7
25
25
21
81
6
20
23
20
22
91
nina 'part/song'
8
15
20
24
67
ya 'flexible, thin'
3
oyla 'string'
1 1
siva 'time'
4 1
1
kavi 'tool'
(2*) 1
1
(1*)
kabisi 'compartment'
1
1
kubila 'land plot'
1
1
Total
11
39
64
67
68
1. CPs produced, pila expected: ke (I, II; 5 tokens), kwe (V; 3 tokens), utu (IV; 1 token), vili (IV; 1 token), ya (IV; 60 tokens). 2. See the comments on dubious (*) CP usage (sec section 3.2.1).
Table 3.2.185. Po'ula 'plantation, grove' Tokens of CP po'ula produced by age group CP expected
I
po'ula 'grove'
II
III
3
18
s a 'nut bunch'
IV
V
Tokens total
9
19
49
2
1
3
1
1
bukwa 'fruit cluster' Total
3
18
11
21
CPs produced, po'ula expected: gula (III; 3 tokens), kwe (I-V; 93 tokens).
Table 3.2.186. Pwanina/pona 'hole, punctured' Tokens of CP pwanina/pona produced by age group CP expected
I
II
III
IV
V
Tokens total
pwanina/pona 'hole'
17
24
36
31
32
140
CPs produced, pwaninalpona expected: kwe (I, IT, TV, V; 23 tokens), nigwa (V; 1 token).
154
Classificatory Particles in Kilivila
Table 3.2.187. Sa 'nut bunch' Tokens of CP sa produced by age group
CP expected
I
II
III
sa 'nut bunch'
1
3
3
V
Tokens total
6
13 13
3
bukwa 'fruit cluster' Total
IV
1
3
3
3
3 6
CPs produced, sa expected: ke (I-V; 1 14 tokens), kwe (I, II, IV, V; 20 tokens), luva* (IV; 3 tokens), Po'ula (IV, V; 3 tokens).
Table 3.2. 188. Si 'small bit' Tokens of CP si produced by age group
CP expected
I
II
III
si 'small bit'
1
ya 'flexible, thin'
4
5
5
Total
5
5
5
IV
V
Tokens total
1 14
1. CPs produced, si expected: bwa (II; 6 tokens), ke (I, II, IV, V; 28 tokens), kwe (II; 3 tokens), utu (I-V; 118 tokens), vili (III-V; 4 tokens). 2. The CP si was produced as a response to the stimulus yoyu 'palm branch' presented to elicit the CP ya; this can be explained by assuming that the consultants thought of 'a small bit of a flexible, thin palm branch'.
Table 3.2.189. Sipu'sheaf Tokens of CP sipu produced by age group
CP expected
I
sipu 'sheaf
II
III
IV
V
Tokens total
11
25
21
31
88
CPs produced, sipu expected: ke (I; 2 tokens), kwe (I-V; 49 tokens), ya (II-IV; 7 tokens).
Table 3.2.190. Sisi"bough' Tokens of CP sisi produced by age group
CP expected
I
sisi 'bough'
7
ya 'flexible, thin'
5
kai 'stone blade' Total
II
III
IV
V
Tokens total
34
33
34
32
140
26
24
30
24
109
2 12
62
(2*) 57
64
1. CPs produced, sisi expected: ke (II, IV, V; 8 tokens), ya (I, III; 6 tokens). 2. See the comments on dubious (*) CP usage (see section 3.2.1).
56
The System of Classificatory Particles in Kilivila
155
Table 3.2.191. Siva 'time' Tokens of CP siva produced by age group CP expected
I
II
III
IV
siva 'time'
V
Tokens total
3
3
CPs produced, siva expected: kwe (I-V; 134 tokens), pila (IV; 1 token), utu (IV, V; 6 tokens), tuta (III-V; 7 tokens).
Table 3.2.192. Suya 'batch of fish' Tokens of CP suya produced by age group CP expected
I
wela 'batch of fish'
II
III
IV
V
Tokens total
6
24
33
27
90
This CP was not expected.
Table 3.2.193. Ta 'basket' Tokens of CP ta produced by age group CP expected ta 'basket'
I
II
III
IV
V
Tokens total
18
34
33
34
32
151
Total
(1*)
1
giwi 'cut'
19
34
33
34
32
1 . CPs produced, ta expected: peta (III-V; 7 tokens), kwe (IV; 1 token), ya (V; 1 token). 2. See the comments on dubious (*) CP usage (see section 3.2.1).
Table 3.2.194. Jam 'sprouting' Tokens of CP torn produced by age group CP expected
I
IV
V
Tokens total
11
8
6
28
6
12
9
34
5
3
8
25
18
II
III
gula 'heap'
3
0 'basketful of yams'
7
uva 'span' Total This CP was not expected.
10
17
156
Classificatory Particles in Kilivila
Table 3.2.195. Tetu 'yams' Tokens of CP tetu produced by age group CP expected
I
II
III
0 'basketful of yams'
IV
V
1
Tokens total 1
This CP was not expected; the CP tetu is a repeater.
Table 3.2.196. To/te 'persons of male gender, human beings' Tokens of CP to/te produced by age group CP expected to/te 'male persons' deli 'group'
I
II
III
IV
V
Tokens total
36 1
36 10
36
36
36
180 11 3
1
yuma 'hand' Total
38
2 46
36
36
38
Consultants 6 and 55 produced the CP to/te as a response to the stimulus presented to elicit the CP yuma; these two consultants took the stimulus yamala 'his/her hand' obviously as a pars pro toto and then referred to 'human beings' in general.
Table 3.2.197. Tuta 'time' Tokens of CP tuta produced by age group CP expected
I
II
siva 'time'
III
IV
V
Tokens total
4
2 1
1
7 1
4
3
1
kala 'day' Total This CP was not expected; the CP tuta is a repeater.
157
The System of Classificatory Particles in Kilivila Table 3.2.198. Utu 'scrap, parts cut off' Tokens of CP utu produced by age group CP expected utu 'scrap' si 'small bit' vili 'untwisted'
I
2 5 1
kabulo 'point'
II
III
IV
V
6 16 2 1
30
30 34 13
19 28 3
22 1
8
7
1
3 5
1
3
kabisi 'compartment' mmwa 'conical bundle'
35 11
siva 'time'
Tokens total
87 118 30 38 4
(4*) 6 (1*) 1
gili 'row'
1
gum 'small piece'
1
kabulu 'sector'
1
1
kavi 'tool'
1
(1*) (2*)
2 4
kubila 'land plot' nina 'song/part'
3
4 3
nunu 'corner/garden'
2 3
(3*)
meila 'song/part'
siwa 'sea portions'
2 1
bubwa 'cut off pila 'part, piece" Total
2
8
32
114
96
2 1
58
1. CPs produced, utu expected: *boma (III; 1 token), bubwa (I, III-V; 8 tokens), bwa (I-V; 20 tokens), ke (I-V; 36 tokens), *luva (V; 3 tokens). 2. See the comments on dubious (*) CP usage (see section 3.2.1). The consultants produced the CP utu as a response to the stimuli presented to elicit the expected CPs kabulo and siva', if we assume that the consultants wanted to emphasize the aspect of 'division' and 'part' that may be an inherent characteristic of 'village sectors' and 'time' in their idiolect we can accept these responses.
Table 3.2.199. Uva 'span' Tokens of CP ova produced by age group CP expected uva 'span'
I
II
III
IV
V
Tokens total
9
6
16
31
CPs produced, uva expected: ke (I-V; 110 tokens), kwe (I, II; 8 tokens), nal* (I; 1 token), tarn (IV, V; 8 tokens).
158
Classificatory Particles in Kilivila
Table 3.2.200. Va 'door, window' Tokens of CP va produced by age group CP expected
I
kaduyo 'entrance'
1
II
V
IV
1
Tokens total 4
3
duya 'door' Total
III
4
6
7
6
10
This CP was not expected.
Table 3.2.201. Vakala 'belt' Tokens of CP vakala produced by age group CP expected
I
II
gili 'row'
III
IV
1
1
V
Tokens total 2
This CP was not expected; the CP vakala is a repeater.
Table 3.2.202. Vaya/vayo 'door, window' Tokens of CP vaya/vayo produced by age group
CP expected
III
IV
duya 'door'
l
1
kaduyo 'entrance'
1
Total
2
I
II
V
Tokens total 2 1
1
This CP was not expected. Table 3.2.203. Vili ' untwisted' Tokens of CP vili produced by age group CP expected vili 'untwisted'
I
II
III
IV
V
Tokens total
1
25
19
23
68
1
2
1
4
si 'small bit' gili 'row'
3
pila 'part, piece'
1
Total
1
26
25
(3*)
1 24
1. CPs produced, vili expected: bubwa* (IV; 2 tokens), bwa (II, V; 2 tokens), ke (I, II, IV, V; 49 tokens), kwe (II, IV, V; 4 tokens), utu (I-V; 30 tokens). 2. See the comments on dubious (*) CP usage (see section 3.2.1).
The System of Classificatory Particles in Kilivila
159
Table 3.2.204. Vosi 'song' Tokens of CP vosi produced by age group CP expected
I
II
meila 'part of a song'
III
IV
V
Tokens total
6
10
2
18
This CP was not expected; the CP vosi is a repeater.
Table 3.2.205. We/a 'botch of fish' Tokens of CP wela produced by age group CP expected
I
wela 'batch of fish'
II
III
IV
V
Tokens total
3
6 8
18
3 4
12 30
14
18
7
oyla 'string of fish'
3
Total
CPs produced, wela expected: na2 (I-V; 60 tokens), suya (II-V; 90 tokens).
Table 3.2.206. Ya 'flexible, thin' Tokens of CP ya produced by age group CP expected ya 'flexible, thin' gili 'row' pila 'part, piece'
II
Ill
IV
V
Tokens total
1
1
12 5
28 23
4 18 11
2 19 9
4 24 12
I
7 8
bukwa 'fruit cluster' duya 'door'
5 4 3
kaduyo 'entrance' lilo 'walk, journey' sipu 'sheaf sisi 'bough'
3
bubu 'cut across'
1
gum 'bit, small piece'
6 1
3
4
12 101 60 14
11 6
9 6
28 17
6
16 7 6
3
3 3 6
4 6 1
yuma 'hand'
1
ta 'basket' Total
22
88
43
54
(1*) 1
66
1. CPs produced, ya expected: ke (I V; 18 tokens), pila (II, IV; 4 tokens), si (I-III; 14 tokens), sisi (I-V; 109 tokens). 2. See the comments on dubious (*) CP usage (see section 3.2.1). The CP ya was also produced as a response to the stimulus presented to elicit the CP lilo; the production of the CP in this context can be justified by assuming that the consultants thought of the rather small paths that connect the villages on the islands.
160
Classificatory Particles in Kilivila
Table 3.2.207. Yam 'day' Tokens of CP yam produced by age group CP expected
I
yam 'day'
II
III
IV
4
8
2
14
3
2
5
11
4
kala 'day' Total
4
V
Tokens total
CP produced, yam expected: kwe (I-V; 141 tokens).
Table 3.2.208. Yam/yama/yuma 'hand' Tokens of CP yam/yama/yuma produced by age group CP expected
I
II
yuma 'hand'
1
2
III
IV
V
Tokens total
2
7
12
CPs produced, yuma expected: kova* (I; 1 token), kwe (I-III; 56 tokens), kweya (III-V; 84 tokens), to/te (I, V; 3 tokens), ya* (IV; 1 token).
Table 3.2.209. Yeglla 'name' Tokens of CP yegila produced by age group CP expected
I
II
iga 'name'
III
IV
V
7
Tokens total 7
This CP was not expected; the CP yegila is a repeater.
Table 3.2.210. Yulai/yule 'bundle of four things' Tokens of CP yulai/yule produced by age group CP expected yulai 'bundle of four things'
I
II
III
IV
6
CPs produced, yulai expected: ke (I, IT; 2 tokens), kwe ( I V ; 148 tokens).
V
Tokens total
3
9
The System of Classificatory Particles in Kilivila
161
Table 3.2.211. Yuva 'shoal' Tokens of CP yuva produced by age group CP expected
I
II
yuva 'shoal'
III
IV
3
V
Tokens total
1
4
CPs produced, yuva expected: buda (I-V; 86 tokens), kasa* (II; 1 token), ke (I; 1 token), kudul* (V; 2 tokens), kwe (I-V; 52 tokens).
Table 3.2.212. O 'basketful of yams' Tokens of CP O produced by age group
CP expected
I
0 'basketful of yams'
II
III 1
IV
V
Tokens total
1
2
CPs produced, 0 expected: kwe (I-V; 117 tokens), tetu (IV; 1 token), tarn (II-V; 34 tokens).
3.2.2.3 CP Types and Tokens Documented in the Corpus of Transcribed Kilivila Speech Up to this point, we have been dealing primarily with data elicited by the questionnaire. In this section, I present the CP types and tokens together with the nouns and word classes with which they were used in my overall corpus of Kilivila speech data. This sample encompasses 34,955 transcribed words. Tables 3.2.216 and 3.2.217 present the relevant CP data; the CPs are listed alphabetically. These 41 types can be ordered according to the number of tokens produced for each type. Table 3.2.218 presents the CPs in this order and gives the percentage of tokens produced for each type in relation to the number of tokens produced for all types (1,564 tokens). With this second set of data, the following CPs were not produced during the elicitation test: bililo 'trip' dumia 'swamplike' kali 'paddle strike' kauya 'creel' lipu 'compartment/creel' na3 'moon, stars' na4 'carving in human likeness' sam 'ginger' yeni 'handful'
4 tokens (consultant V17) 1 token (consultant 12/10) 1 token (consultant V14) 4 tokens (consultant IV3/33) 4 tokens (consultant VI5) 5 tokens (consultant IV3/33) 4 tokens (consultant VI1) 3 tokens (consultant V11) 7 tokens (not assignable to individual speakers) I token (consultant V17)
We will now turn to the questions raised in section 3.1.
Table 3.2.213. CPs produced although not expected Tokens in each age group CP produced
Tokens
CP expected
beku 'stone blade' (*)boma '???' bwalita 'sea' doba 'skirt' guba 'bundle/taro'
2 1
kai 'stone blade' utu 'scrap'
6 1
*iki '???' ka'i 'tooth'
1
siwa 'sea portions' kudul 'band of fibres' mmwa 'conical bundle' nina 'part of a song' kudu2 'tooth' oyla 'string' kwoila 'clay pot' kapwa 'bundle' nina 'part of a song' pwanina 'hole'
3 124
(*)ka'i 'tooth' (*) kava '???'
3 1
luba 'bundle' megwa 'magic' nigwa 'hole'
77 12 1
1
II
III 1
IV
V
1
1 2 1
4 3
1 13
27
28
26
9
21
25
1
5
6
30
3 1 1
21 1
peta 'basket'
7
ta 'basket'
3
1
3
6
24
33
8
6
suya 'batch/fish'
90
wela 'batch of fish'
tarn 'sprouting'
28
gula 'heap'
8 34
uva 'span'
tarn 'sprouting' tarn 'sprouting'
1
0 'basketful of yams'
tuta 'time'
7
siva 'time'
tuta 'time'
1
kala 'day'
va 'door'
10 4
5
3
6
12 1
9
4 1
2
1
6
4
duya 'door' kaduyo 'entrance'
11 7
0 'basketful of yams'
tetu 'yams'
va 'door'
3
1
3
vakala 'belt'
2
gili 'row'
1
1
vaya/vayo 'door'
2 1
duya 'door'
1
1
vayalvayo 'door' vosi 'song' yegila 'name' Total
18 7
27
1
kaduyo 'entrance' meila 'part of a song'
6
iga 'name'
7 15
67
119
10
2
147
104
Table 3.2.214. Dubious cases of CP production Tokens by age group CP expected bubu
CP produced (dubious usage) buda gula kudul nal
1
II
III
1 2
3 1
2
luva
duya
nal
gili
utu vili
1
kasa na2 ta luva
3 1
3 3
\
3 1
buda kasa nal
3
iga
kasa
1
kabisi
nunu
kabulol
bubwa
1
kabulu2
kada
3
gum
kai kavi
nal sisi pila
V
1 1
bukwa
giwi
IV
3
1
1 2 1
utu kubila
utu
kweya
nal
kwoila
kava
lilo
kada
mmwa
luva utu
nina
en
1 2 1
1 1 3 1
kabulo2
oyla
ka'i pila
3
1
iki
nunu
3
3 3 1
1
3
sa
luva
siwa
utu
3
utu
luva boma
1
uva
vili
nal bubwa
vilo
kada
yamalyuma
kova ya
yuva
kasa kudul
3 1
2 2 1 1 1 2
Total
9
32
19
17
11
Total not acceptable
8
22
14
3
7
Total dubious/acceptable
1
10
5
14
4
166
Classificatory Particles in Kilivila
Table 3.2.215. Dubious cases of CP production by gender of consultants Age group I II III IV V
Total
Females
Males
8 16 15 10 4
16 4 7 7
9 32 19 17 11
53 (= 60%)
35 (= 40%)
88 (= 100%)
All
1
Table 3.2.216. CP types and tokens in the corpus of Kilivila speech data Tokens/CP type
Noun(s) CP refers to
4/bililo 'trip' 5/bubwa 'cut'
kova 'fire', kebudaka 'beam', kelagim 'beam', tou 'sugar cane'
18/buda 'group'
, kukumatua 'bachelor', boda 'group', gugwadi 'children', vivila 'girl'
6/bukwa 'cluster'
nuya 'coconut', bwebwai 'green coconut'
2/bwa 'tree'
kebaxi 'kind of wood'
1/dumia 'swamp' 4/gudi 'child'
vivila 'girl', gwadi 'child'
4/gula 'heap' 5/kabulu 'sector'
mani 'money', kena 'spatula'
10/kada 'road'
koya 'mountain' (but: repaired)
22/kai 'blade'
regisa 'axe', kema 'axe', ligogu 'adze' (+ kinship term produced in several mourning formulae)
valu 'village', kabulu 'village sector' vaya 'river', keda 'road'
I/kali 'paddle' 9/kasa 'row'
buna 'cowrie shell', kai 'tree', tauwau 'men', vivila 'girl'
8/kauya 'creel' 236/Ke 'wood'
kai 'tree, tree and fruit names', poseni 'poison', keyala 'spear', gebaku 'bow', kaimili 'mortar', kaboma 'plate', yaguma 'lime pot', kenuya 'dish', sinata 'comb', waga 'canoe', selinboti 'sailing boat', dingi 'dinghy', kemolu 'canoe', riga 'outrigger', lamina 'outrigger', lagim 'canoe board', sukmaku 'top on a canoe prow', kivaya 'fish trap', vadola 'opening of a basket', vilivili 'body of a basket', medisini 'medicine', pent 'pen', youdila 'tool', kevalapu 'rafter', kavala 'perlin', kivi 'rafter', dagala 'stick used as support', budaka 'plank', liku 'foodhouse', kaitukwa 'walking stick', tataba 'carved board', boi 'symbol in a canoe prow', koni 'sign of honor', kahu/ulu 'point', xalihu 'mirror', kuvi 'type of yams', pwaka 'lime', tobaki 'tobacco', ka'eki 'shell', dakuna 'stone', bweta 'wreath of flowers', kova 'fire', kenuba 'firewood', bela 'drum', paledi 'plate', kukwanebu 'story' (but: repaired!), serota 'platform', lubakatakela 'rainbow', keyoyova 'plane', tuta kevau 'time for launching'
The System of Classificatory Particles in Kilivila
167
Table 3.2.216. (continued) Tokens/CP type
Noun(s) CP refers to
6/kova 'fire'
kova fire'
3/kudu 'fibers'
doba 'skirt', vana 'decoration, herbs'
1lkubila 'land' 361/kwe 'thing'
vavagi 'thing', gugua 'goods', valu 'village', PNG 'Papua New Guinea', simla 'island', koya 'mountain', yam 'day', wiki 'week', tctu 'year', manakapu 'reef, bwala 'house', kehila 'veranda', kwadeva 'beach', pwepwaya 'ground', bagula 'garden', hugulela 'plants in the garden', kuvi 'yams', tctu 'yams', salutu 'yams', kaula 'yams', kuku 'tobacco', vadila 'pandanus', kauveluva 'food, not yams', kwetutum 'your knee', kunula 'her/his hair', pwaneta 'skull', puvala 'his testicles', matala 'her/his eyes', bulaga 'shadow', boda 'group', higa 'word', bigatona 'speech', kweatin 'question', liliu 'myth', livala 'talk', livalela 'news' nanam'sa 'idea', kukwanebu 'story', katupoi 'question', megwa 'magic', bomala 'tabu', sula 'mistake', mokita 'truth', kabitam 'skill', mitin 'meeting', mwasawa 'fun', kove.ia 'competition', pilasi 'help', paisewa 'work', agutoki 'thanks', bubunela 'custom', kato'ula 'illness', tauya 'triton', buna 'cowrie shell', kweduya 'shell', poseni 'poison', dakuna 'stone', youdila 'tool', kehudaka 'beam', vatila 'construction to transport something', vatunu 'rope', nabala 'bottom of a basket', kena 'spatula', kwena 'pot', moi 'mat', doba 'skirt', tarapwapwa 'carved symbol', katukwaka 'cave', sopi 'water', ginigini 'writing', stoa 'store', kwestin 'question', edyukesen 'education', beta 'bell', lekodi 'record', mani 'money', laita 'light', tepi 'tape', proyecta 'projector', ensini 'engine', hour 'hour', namha 'number', ximenti 'cement', mamwa 'hour', niva'ila 'calm sea', mokolu 'tree'
8/kwela 'pot'
kwena 'clay pot', kerosina 'drum full of kerosine'
2/kweya 'limb'
kaikela 'her/his leg'
6/liku 'compartment'
kemolu 'small canoe', liu 'beams connecting the outrigger with the canoe', valota 'stick connecting outrigger and outrigger platform'
5/lipu 'compartment of a creel' 174/nal 'female'
85/na2 'animals'
(names of several women), vivila 'girl', numwaya 'old woman', sinehada 'lady', kwava 'wife', gwadi 'child', 'kinship terms', bwagau 'sorceress', lokwam 'cannibal', dokonikani 'man-eating monster', beila 'monster' , mauna 'animal, name of specific animals', yena 'fish', tabuya 'canoe prow'
4/na3 'moon'
tuhukona 'moon'
3/na4 'carving'
tokwalu 'carved figure on a canoe board'
36/pila 'part'
, kukwanebu 'story', vinavina 'verses', sopa 'joke', wosi 'song', megwa 'magic', tegala 'her/his ear', kekwahula 'drawing', paxporti 'passport'
Classificatory Particles in Kilivila
168
Table 3.2.216. (continued) Tokens/CP type
Noun(s) CP refers to
1/pwana 'hole'
moropu 'hole'
7/sam 'ginger'
neya 'ginger'
1/sisi 'bough'
revaya 'name of a tree'
8/ta 'basket'
peta 'basket', taboda 'door' kuvi 'type of yams'
1/tam 'sprouting' 437/to/te 'human, male'
— (names of several men), tau 'man', tauwau 'men', tomwaya 'old man', vivila 'girl', nunumwaya 'old woman', gwadi 'child', gugwadi 'children', tommota 'people' (corpse, kinship terms, forms of address), pikisila 'picture', gumanuma 'Europeans', dimdim 'Europeans', doketa 'doctor', toliwaga 'owner of a canoe', referi 'referee', kini 'king', dokonikani 'man-eating monster', tobiu 'referee'
2/utu 'scrap'
, unasu 'small yams'
7/uva 'span'
, peta 'basket', kivaya 'fish trap', nuya 'coconut'
66/ya 'flexible'
bologu 'tree leaves', sasova 'banana leaves', yakvesi 'banana leaves', kum 'bread fruit', keibwibwi 'pandanus', bistta 'pandanus streamer', veyugwa 'tree leaves', kwega 'tree leaves', yagavana 'tree leaves', peta 'basket', kivaya 'fish trap', kauya 'shoulder bag', vatunu 'rope', yuvaiyo'ula 'rope, name of several knots', naya 'sail', pwepu 'paper', leta 'letter', tau 'man' (metaphor!)
1/yeni 'handful'
pwaka 'lime'
1/yule 'bundle' 1/yuma 'length'
pepe'u 'small basket'
2/0 'basketful of yams' Total: 41 types; 1,564 tokens The note "(but: repaired)" refers to the fact that a speaker first produced this CP with the noun but then repaired the inadequate or ungrammatical phrase, replacing the inappropriate CP with a grammatically correct CP.
Table 3.2.217. CP type and tokens—word classes (corpus of transcribed Kilivila speech) CP type
Demonstrative pronoun
bililo 'trip' bubwa 'cut off
1
buda 'group'
6 4
bukwa 'cluster'
Numeral
Adjective
4 4 7
5 2
Interrogative pronoun
Total
4 5 18 6
169
The System of Classificatory Particles in Kilivila Table 3.2.217. (continued) CP type
Demonstrative pronoun
bwa 'tree' dumia 'swamp' 1 4
kadalkeda 'road' kai 'stone blade'
8
2
6
1 2 1
49
kova 'fire' kudu 'fibers'
6 1
2
kwabila 'land' kwe 'thing' kwela 'pot'
1
liku 'compartment' lipu 'compartment' nal 'female' na2 'animals' na3 'moon' na4 'carving' pila 'part' pwanina 'hole' sam 'ginger' sisi 'bough' ta 'basket' tam 'sprouting' tolte 'male' utu 'scrap' uva 'span' ya 'flexible' yeni 'handful' yule 'bundle' yuma 'length' 0 'basketful of yams' Total
4 4 5
1 14
10 22 1 9
6
8 47
2
236 6 3 1
1
361 8 2
167 3 1
114
79
1 1
4
2
4
6
5 136 69 1
6 13
31 3
1
3 8
4
36 1
7 1 1
7 1
4
3
110
70
5
7
2
8 1
1 252
34
2 7 23 1 1 1 2
874
5 174 85 4
3
3 24 1
Total 2 1
1 1
9 2 138
kweya 'limb'
Interrogative pronoun
5
kali 'paddle' kasa 'row' ke 'wood'
Adjective
2
gudi 'child' gula 'heap' kabulu 'sector'
kauya 'creel'
Numeral
407
437 2 7 66 1 1 1 2
272
11
1,564
170
Classificatory Particles in Kilivila
Table 3.2.218. Percentage of tokens for each CP type in corpus of Kilivila speech CP type tolte 'male, human' kwe 'thing' ke 'wood' nal 'female' na2 'animal' ya 'flexible' pila 'part' kai 'stone blade' buda 'group' kada 'road' kasa 'row' kauya 'creel' kwela 'pot' to 'basket' sam 'ginger' uva 'span' bukwa 'cluster' kova 'fire' liku 'compartment' bubwa 'cut off kabulu 'sector' lipu 'compartment' bililo 'trip' gudi 'child' gula 'heap' na3 'moon' kudu 'fibers' na4 'carving' bwa 'tree' kweya 'limb' utu 'scrap' 0 'basketful of yams' dumia 'swamp' kali 'paddle strike' kwabila 'land plot' pwanina 'hole' sisi 'bough' tarn 'sprouting' yeni 'handful' yule 'bundle' yuma 'length'
Tokens
437 361 236 174 85 66 36 22 18 10 9 8 8 8 7 7 6 6 6 5 5 5 4 4 4 4 3 3 2 2 2 2 1 1 1 1 1 1 1 1 1
Percent (100% = 1,564 tokens)
28 23 15 11 5 4 2 1 1 0.6 0.6 0.5 0.5 0.5 0.4 0.4 0.4 0.4 0.4 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.1 0.1 0.1 0.1 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06
The System of Classificatory Particles in Kilivila
171
3.3 RESULTS In this section, I try to answer the questions raised in section 3.1. The headings within this section repeat the questions as they were stated in section 3.1. 3.3.1 Which Formulatives Constitute the Kilivila CP System and What is the Actual Occurrence of CPs in Recorded Tests? Before I can start listing the formatives that constitute the Kilivila CP system, I first have to classify the CPs given in section 3.2. From the beginning, I have emphasized that I only provisionally assigned the status "CP type" to all the CPs listed. Now, with all the data presented, I will further refine this rather broad and preliminary concept. The reasons that force me to do this are obvious. The attentive reader of the tables presented in section 3.2 will notice that the label "CP type" was assigned not only to CPs like kai, bogi, and kova, but also to CPs like buda, kada, kweya, and kwoila, which were listed with variants (boda, budu', keda; kwaya, keya; and kwela, respectively). To arrive at a consistent description and nomenclature, the label "CP type (proper)" should subsume all variants of a CP type. On the basis of this consideration, the following modifications in the definition of CP type are necessary: 1. The formatives iga 'name' and yegila 'name' are to be regarded as two variants of one CP type (see Tables 3.2.18, 3.2.56, 3.2.120, 3.2.145, and 3.2.209). 2. The formatives duya 'door, entrance' and kaduya, kaduyo 'entrance, door' are to be regarded as three variants of one CP type (see Tables 3.2.11, 3.2.18, 3.2.49, 3.2.62, 3.2.138, and 3.2.151). 3. The formatives peta 'basket' and ta 'basket' are to be regarded as two variants of one CP type (see Tables 3.2.18, 3.2.25, 3.2.32, 3.2.49, 3.2.104, 3.2.183, and 3.2.193). 4. The CPs va 'door' and vaya, vayo 'door' are to be regarded as three variants of one CP type (see Tables 3.2.111, 3.2.113, 3.2.200, and 3.2.202). The reasons for these redefinitions are obvious in the cases of iga/yegila and peta/ta. Yegila is a repeater, and iga is a classifier, probably the shortened form of the repeater. The same holds true for peta and ta. Here, peta is the repeater and ta is the classifier (again, a shortened form of the repeater). Similarly, I can justify handling the CP variants duya/kaduya/kaduyo and va/vaya/vayo in this way. Although only the formative vaya is a repeater (at least to a certain degree), duya can be regarded as a shortened form of kaduya and va as a shortened form of vaya. As stated elsewhere (G. Senft 1986:242, 219, 400, 407), the Trobrianders do not differentiate between the formatives duya and kaduyo and va and vaya lexically and semantically. The meaning of va includes the meaning of duya. Thus, we might also regard the formatives va, vaya, vayo, duya, kaduya, kaduyo as variants of only one
172
Classificatory Particles in Kilivila
CP type (see also Tables 3.2.4, 3.2.11, 3.2.18, 3.2.25, 3.2.138, 3.2.151, 3.2.200, 3.2.202). However, this proceeding also implies a chain of development that leads from the formatives kaduya/duya and vaya/vaya to the classifier va. Although this argument is plausible, it is highly speculative. Therefore, having hinted at such a possible interpretation of the data, I will continue to differentiate between the CP type va/vaya/vayo, on the one hand, and the CP type duya/kaduya/kaduyo, on the other hand. In the Tables of section 3.2, I generally listed polysemous CPs like kwe and ke without any differentiation of their various meanings. However, I differentiated between the meanings of some CP types like na (nal-na4). Again, for the sake of a consistent description and nomenclature, the following modifications with respect to the definition of the concept CP type are necessary: 1. Kabulo/kabulu is to be regarded as one polysemous CP. 2. Kudu is to be regarded as one polysemous CP. 3. Na is to be regarded as one polysemous CP. According to my consultants, the formatives *boma, *iki, and *kava cannot be accepted as belonging to the system of Kilivila CPs. If I take all these considerations into account, I can now list the CP types that constitute the elicited, observed, and documented system of CPs in Kilivila. Table 3.3.1 presents these CP types together with their meanings. Some of these meanings are not documented in my corpus of transcribed Kilivila speech data, but are the result of my lexicographic work. I should mention here that during my restudy on the Trobiand Islands in 1989, my consultants produced three additional CPs, namely, num. 'magic, magical formula' (consultant: Vadomna); tili 'bits of lime clinging to a lime spatula' (consultant Vadomna); sebulu 'grass skirt for little girls' (consultants: Ilakelava and Kadawaya). However, these CPs seem to be either very rarely used or almost obsolete, and they are mentioned here only for the sake of completeness. I want to emphasize that it is the 88 classifiers presented in Table 3.3.1 that are the focus of attention of the analyses presented here. As stated in section 3.1, a complete description of the CPs should include the formulation of the rules speakers use in their production of a certain CP. This description is given in section 3.3.4.2. For the purposes of the present section, the mere list of CPs that constitute the elicited, observed, and documented (sub)system of CPs in Kilivila should suffice. For the sake of completeness, a list of the CPs and their glosses that are mentioned in Lawton (1980:213-256), but were not produced by my consultants, are presented in Table 3.3.2. Capell (1969:60-62) lists three formatives that have not been mentioned so far: gubwa, ilu, and ukdu. My consultants rejected Capell's gloss 'square' for gubwa, replacing it with 'group of four'; however, they emphasized that
The System of Classificatory Particles in Kilivila Table 3.3.1. Kilivila CP types CP type llbeku
Gloss(es) and comments
2/bililo
'stone blade' 'trip't
3/bogi
'night'
4/bubulbobu/bobo 5/bubwa
'cut across, cut transversely, (block) cut off 'cut across, part(s) cut off
6/buda/bodalbudu
'group, team, crowd'
7/bukwalbuko
a. 'fruit cluster' b. 'cowries tied into a specific cluster'T
8/bwa
'trees, wooden things'
9/bwalita 10/deli
'sea' 'company, group on the move'
11/doba
'skirt made of banana leaves, grass skirt'
12/duli
'cluster, bundle'
12/dumia
'swamp, swamplike'
14/duya/duyo/kaduya/kaduyo
'door, entrance'
15/gili
'row' 'cut'
16/giwi 17/guba
'bundles of taro'
18/gudi
a. 'child'; b. 'immature human'
19/gula/guli/gulo/guno
'heap, group'
20/gum
'bit, small piece'
2l/iga/yegila
'name'
22/kabisi
'compartment of a foodhouse, section/division in a foodhouse'
23/kabulo/kabulu
a. b. c. d.
24/kada/keda
a. 'road, track' b. 'way in which something is done'
25/kai
'stone blade' 'tooth'
26/ka'i 27/kala 28/kali
'protuberances' 'village sectors; areas of authority' 'cape, point, peninsula' 'half of something'
'day' 'paddle strike'
29/kapwa/kapo
a. 'bundles (wrapped up), parcel' b. 'nest of birds'
30/kasa
'row, line'
This meaning of the CP type was not elicited in the lest. This (these) meaning(s) of the CP type are the result of my lexicographic research.
173
174
Classificatory Particles in Kilivila
Table 3.3.1. (continued) CP type
Gloss(es) and comments
3l/kauya
'fish trap, creel'
32/kavi
'tool'
33/ke
a. b. c. d.
'wooden things' 'rigid, long objects' 'unmarked form for inanimates' (general classifier) fire'
34/kila
'clusters/hands of bananas'
35/kova
'fire, fireplace'
36/kubila/kwabila
'large land plot'
37/kudu
a. 'band of fibers' (especially the band of fibers at the waistband of a grass skirt) b. 'tooth' c. 'bundles of lashing creeper'
38/kumla
'earth oven'
39/kwe
a. 'thing, anything indefinite or unknown, unmarked form for inanimates' (general classifier) b. 'shells and clams'
40/kweya/kwaya/keya
a. 'limb, severed limb' b. 'yard (land adjacent to a house)
41 Ikwoilalkwelalkwaylkwailal kweikwalkwena
'clay pot, potlike'
42/kwoya/koya
'mountain, hill'
43/liku
a. 'compartments of a foodhouse, compartments of a canoe' b. 'area of authority'
44/lila
'bough, branch, leaf
45/ltio/lola/lilo'u
a. 'walk, journey' b. 'number of times going somewhere' c. 'number of times doing something'
46/lipu
'compartment of a creel, tier'
47/luba 48/luva
'bundle (of rolls), parcels (of taro pudding)' a. 'wooden dishes' (kaboma-type, full of one's share of food during a food-distribution ceremony/ritual) b. 'tied bundle'
49/megwa
'magic, magical formula'
50/meila/mavila
a. 'part of a song, part of a magical formula' b. 'part of a (Bible) chapter' c. 'part of a day'
5l/mmwa/mmo
'conical bundle (of taro)'
52/na
a. b. c. d. e. f.
'persons of female gender' 'animals' 'stars, planets, moon' 'carvings in human likeness' 'corpses' 'spirits, dwarfs'
The System of Classificatory Particles in Kilivila
175
Table 3.3.1. (continued) CP type
Gloss(es) and comments
53/nigwa/nigo
a. b. a. b.
54/nina
'hole' 'nest' 'parts of a song' 'idea, thought'
55Inutulnotu
'kneaded things, dot, drop'
56/nunu
'corner(s) of a garden'
57/oyla
a. 'string' b. 'fish on strings'
58/petalta
a. 'basket' b. 'contents of a basket' (but not baskets of yams!)
59/pila/pa
'part, piece'
60/po'ula
a. 'plantation, grove' b. 'heap, group'
61/pwaninalponalponinalponul polulpwana
'punctured, something with a hole in it, hole'
62/sa
'nut bunch'
63/sam
'ginger' (in play accompanying verses)
64/si
'small bit'
65/sipu
'sheaf (Lawton (1980) gives the additional meanings 'tangle, tangled line, rope, net, string')
66/sisi
a. 'bough' b. 'cut off part of a tree' c. 'division of a magical formula'
67/siva
a. 'time' b. 'number of times doing something'
68/siwa
'sea portions, ownership division with reference to fishing rights'
69/suya/suye
'batch of fish on strings'
70/tam
'sprouting, sprouting yams'
71/tetu
'yams'
72/to/te
a. 'persons of male sex' b. 'human beings'
73/tuta/tuto
'time, occasion'
74/utu
'scrap, parts (cut off), small particles, fragments'
75/uva
a. 'span, measure (the span of two extended arms, from tip to tip)' b. 'items measured in spans'
76/va/vaya/vayo/vata
a. 'door, window' b. 'river, creek, sea passage'
77/vakala
'belt of spondylus shell discs'
78/vili 79/vi/o
'untwisted' 'place, area, village'
80/vosilwosi
'song, parts of a song'
176
Classificatory Particles in Kilivila
Table 3.3.1. (continued) CP type
Gloss(es) and comments
81/wela
'batch of fish, string of fish'
82/ya
'flexible things, thin things'
83/yam
a. 'day' b. 'number of days'
84/yumalyamlyumalyama
a. 'hand' b. length, measure (the span of two extended arms, from the fingertips of one hand to the wrist of the other hand)' c. 'yard' (measurement)
85/yeni
'a handful of something'
86/yulai/yule
'bundle of four things'
87/yuvalyuwo
'shoal'
88/0
'a basketful of yams' (this zero classifier is only used when baskets of yams are counted)
Table 3.3.2. CP types from Lawton 1980 that were not listed in Table 3.3.1 CP type
Gloss
Ll/biga
'word, statement, message, public speech'
L2/biko
'coconut bunch'
L3/bili
'roll, anything rolled up'
LA/bo
'cut across'
L5/bubulo
'anything manufactured or created'
L6/bukolbuku
'buried'
L7/buliga
'floor or story'; 'drawers or shelves in series'; 'horizontal divisions in a foodhouse'
L8/bulu
'half-submerged (boat)'
L9/buluwo
'ten-group (animals)'
L10/dala/dila
'family line within the clan group'
L11/dodiga
'load, contents'
L12/gabu/gubu
'fireplace; place where fire or sparks have burned body; batch of roasted food'
L13/gibu
'sufficient'
Ll4/gini
'mouthful of food'
L15/givi
'serve of fish'
Ll6/gubo/gubu
'garden division'
L17/gugulo
'heap; meeting, gathering'
L18/gulo
'group; heap; bundle of fibres laid side by side'
The System of Classificatory Particles in Kilivila
177
Table 3.3.2. (continued) CP type
Gloss
L19/ligi
'wind'
L20/ika
'tens of things'
L21/liwo
'group'
L22/kabila
'large cut/serving of meat, fish, etc.'
L23/kabilikova
'fireplace'
L24/kadida
'very small garden division; division of a task between several workers'
L25/kaiga/kaigi
'sound of a voice'
L26/kailiku
'suburb, part of a village'
L27/kaiyuvai/yuvai
'layers, rows of things'
L28/kalipo
'site, part selected for some purpose'
L29/kalivisi/kaluvisi
'large garden division'
L30/kalo
'two-bundle crustacean'
L3l/kaluwo
'ten days; ten groups of kai, ke items'
L32/kapu
'mouthful of drink'
L33/kapuli
'group of parcels, cargo of goods, load of people'
L34/kapupu
'grove tuft of hair left on head after head shaved'
L35/kasila
'ten group (wealth)'
L36/katukuni
'reel, one turn in a roll of anything'
L37lkatuluwo
'large group'
L38/katupo
'section, quarter'
L39/kaulo
'ten group (strings of fish)'
L40/kaya
'half (piece of food)'
L41lkevala
'batch drying'
L42/kipu
'cut of meat, mouthful of flesh'
L43/kumila
'clan'
L44/kunolkuna
'rain'
L45/kununu
'serving of greens, number of fibres laid together'
L46/kupa
'line rolled in loose bundles; serving of greens (uncooked)'
L47//kupo
'two-string (fish, marine creatures)'
L48/kuwo
'crumb'
L49/kwailuwo
'tens of things'
L50/lada
'small fishing spot'
L51/lapou
'a third of
L52/ligila
'group action, group of acts completed'
L53/lilivi L54/livisi
'forked stick'
L55/lukuva
'growing bundle'
L56/lupo
'very small garden division'
L57/miga
'appearance, face'
'shelf, divisions in a foodhouse, contents of one division'
178
Classificatory Particles in Kilivila
Table 3.3.2. (continued) CP type
Gloss
L58/moya
'limb; position in family line'
L59/mweli
'practices; bundle of leaves'
L60/nakwa
'thing'
L61/no
'blow'
L62/puli
'bunch'
L63/pulu
'garden mound'
L64/pupai
'layer of filth'
L65/pwa
'excrement'
L66/pwasa
'rotten'
L67/sega
'branching'
L68/seluva
'bundle being tied'
L69/seuyo
'lagoon'
L70/sisili
'cut of meat'
L71/siyolsuyo
'things strung through hole'
L72/sobulo
'growing (shoot)'
L73/soulo
'fishing spot'
L74/tabili
'roll'
L75/tabudo
'room'
L76/tavi
'loose coil'
L77/teni
'tight coil'
L78/tubo
'generation'
L79/tupila
'fleet'
L80/udi/udila
'land tract'
L81/umila
'grove' 'two bundle'
L82/uwo L83/vala
'small garden division'
L84/wouyo
'newness'
L85/yivi
'serving of food pieces'
this CP is hardly ever used. My consultants did not recognize the CPs ilu and ukdu; it should be mentioned here that Capell's formative ukdu does not agree with Kilivila syllable patterns (G. Senft 1986:20-122). Thus, if we add to the 88 CPs in Table 3.3.1 the three newly found CPs num, sebulu, and tili, as well as the 85 CPs mentioned by Lawton (1980) and the CP gubwa quoted by Capell (and confirmed by my consultants), we arrive at an overall CP inventory of 177 formatives.2 However, I want to emphasize 2 The discrepancy between the number of CPs given here and in G. Senft (1986:69-75) is due to considerations similar to those presented in the introductory paragraphs of this section.
The System of Classificatory Particles in Kilivila
179
Table 3.3.3. Percent of CP tokens in relation to all transcribed words in overall corous of Kilivila sneech data Word class
No. of CPtokens
Percent
2.5%
Demonstrative pronouns
874
Numerals
407
1.2%
Adjectives
272 11
0.8% 0.03%
1,564
4.53%
Interrogative pronouns/adverbs Total t34,955 words = 100%.
once more that only the 88 CPs presented in Table 3.1.1 are used in the research presented in this book. I will now address the issue of the actual occurrence of CPs in recorded texts. To do this, I must refer to the data in appendix B, to the discussion of the CP types and tokens produced in my overall corpus of Kilivila language given in section 3.2.1, and especially to Tables 3.2.216 and 3.2.217. As I mentioned in section 3.2.1, the production of CPs in actual communication is highly dependent on the text category. The text category influences not only the number of CP tokens found with demonstrative pronouns, numerals, adjectives, or interrogative pronouns/adverbs, but also the number of different CP types produced by the consultants. With the different text categories produced by the consultants in the second data set, I cannot justify any inferences from CP production to characteristics of the individual consultants' speech behavior. Thus, it remains an open question whether the text category or the individual consultant's characteristics of speech behavior, or both, are responsible for the CP production observed in this data set. As stated in section 3.2.1, I only dare to assign to the observations and interpretations presented there the status of anecdotal findings. Nevertheless, as far as I know, Tables 3.2.216 and 3.2.217 present for the first time in the published literature a detailed summation of classifier types and tokens from actual speech (and, in the Kilivila case, with all four word classes that use CPs as a means of their word formation). On the basis of the data presented in these tables, I can answer the question about the actual occurrence of CPs in recorded texts, keeping in mind the precautions mentioned previously. My corpus of transcribed Kilivila speech data consists of 34,955 words,3 including 1,564 CP tokens representing 41 different CP types. If I relate the number of CP tokens to the total number of words, I find that words using CPs in their word formation represent 4.5% of all words documented in the recorded and transcribed texts. Table 3.3.3 gives the exact percentages of the actual occurrence of CPs for all four word classes. 3 The definition of "word" is based on G. Senft (1986:28-102); however, affixes are not counted separately here.
180
Classificatory Particles in Kilivila
Although we lack similar data from other classifier languages, these data may serve as a guideline for future discussions about the actual occurrence of CPs in recorded texts. Whether these numbers should be interpreted as being relatively low is a question that remains open. 3.3.2 How Is the CP System Acquired by Trobriand Children? As stated in section 3.1, this question comprises a number of different aspects. I will try to answer these separate subquestions in the same order that they were raised in section 3.1. 3.3.2.1 What Is the General Temporal Progress of the CP Acquisition Process? In section 3.1,1 mentioned that the informal pretests I carried out revealed that children start to acquire the CP system at the age of approximately 4 (+ 6 months). This result of the informal pretests is confirmed by the data presented in section 3.2. As the first analyses and interpretations of the data presented in sections 3.2.1.1-3.2.1.3 show, I can record the following results with respect to the temporal progress of the childrens' CP acquisition documented in their CP production. 1. Children start to acquire and to produce CPs at the age of 4 ( + 6 months). 2. Five years is the critical age with respect to the CP acquisition process. 3. At about 6 years of age, there is a kind of "explosion" in the production of CPs (see Tables 3.2.1 and 3.2.2). 4. Nine years is another critical age with respect to the CP acquisition process: the frequency of CPs produced (in the elicitation test) increases rather dramatically at this age (see Tables 3.2.8-3.2.10). The 'rule of thumb' generally accepted in psycholinguistic acquisition research, quoted in section 3.2.1.3 in Klein's formulation (Klein 1984:21) is confirmed by our data. If we compare the data produced by the consultants in age group III with that produced by the consultants in age groups I and II, on the one hand, and with the data produced by the consultants of the age groups IV and V, on the other hand, it is clear that, on the whole, the CP acquisition process ends during or after puberty. The 12 consultants in age group III produced 109 CP tokens more than the consultants in age group II; however, they only produced 1 token less than the consultants in age group V and 31 CP tokens less than the consultants in the age group IV (see Tables 3.2.15, 3.2.8, 3.2.22, and 3.2.29). Moreover, the consultants in age group ITT showed the highest mean value of different CP types produced of all five age groups (see Table 3.2.124). Although this value exceeds the respective means for age groups IV and V by 2 and 1 points only, it surpasses the value for age group II by 12 points.
The System of Classificatory Particles in Kilivila
181
3.3.2.2 Are There Any Differences in the CP Acquisition Process With Respect to the Production of Demonstrative Pronouns, Adjectives, and Numerals That Use CPs as Morphemes in Their Word Formation? I can answer this question by summarizing the first analyses and interpretations of the data presented in sections 3.2.1.1-3.2.1.5. The results are based on the data presented in Tables 3.2.1, 3.2.8, 3.2.15, 3.2.22, and 3.2.29. If we compare the data presented there, we note the following: 1. Children who range in age from 4 to 7 years produce CPs most often with adjectives, then with numerals, and then with demonstrative pronouns. I assume that the different CP production frequencies documented in Table 3.2.1 can be interpreted as indices for the sequential order of the acquisition (and production) of the respective word classes. This results mirrors the general tendency documented in the elicited CP data. It neglects the specific use of CPs with the three word classes that depends on certain documented text categories in my overall corpus of Kilivila speech data (see section 3.2.1.1 and the comments on Table 3.2.6). 2. Nine years is not only a critical age with respect to the rather dramatic increase of CPs produced, it is also a critical age with respect to a balanced usage of CPs with adjectives, numerals, and demonstrative pronouns. This result again neglects the specific use of CPs in certain (documented) text categories (see section 3.2.1.2 and the comments on Table 3.2.13). The data gathered during the CP elicitation test show a balanced use of CPs with the three word classes elicited for all consultants in age groups III, IV, and V. For the sake of completeness, it should be noted that the data documented in the overall corpus of Kilivila speech show that the actual occurrence in recorded texts of the interrogative pronoun/adverb that uses CPs in its word formation is rather low indeed (see Tables 3.2.14, 3.2.21, 3.2.28, 3.2.35, and 3.3.3). Finally, I want to mention here once more the rather interesting interpretation provoked by the first analyses of the data presented in Table 3.2.6 (see also Table 3.2.13). These data show a rather high proportion of demonstrative pronouns in recorded texts produced by children from 4 to 7 years old. Because these pronouns were produced within the text category "fairy tale," I arrived at the following interpretation: It may be that fairy tales are told not only to enterain children and transmit cultural ideas and values, but also as a means to foster the acquisition process with respect to certain grammatical features—here demonstrative pronouns—that are obviously difficult to acquire by children of this age.
182
Classificatory Particles in Kilivila
3.3.2.3 Are There Any Gender-Specific Differences in the CP Acquisition Process? I can answer this question by summarizing and elaborating on the first interpretations and analyses presented in section 3.2, especially in section 3.2.1. Thus, I note the following: Children ranging from 4 to 7 years old show rather distinct gender-specific differences in the CP acquisition process (see Table 3.2.1). Boys produce more CP types (4 on the average) and more CP tokens (130) than girls. Moreover, the order of acquisition and production of CPs with first adjectives, then numerals, and then demonstrative prounouns is more marked in girls than in boys. These results obviously hold for children who are 8 years old, too, as the data produced by consultant 14 in age group II (see Table 3.2.8) indicate. However, these gender-specific differences become less marked if we look at the speech behavior of children ranging from 9 to 14 years of age. I emphasized before that 9 years is another critical age in the CP acquisition process. Thus, I had to exclude consultant 14, Olopola, who had just passed her eighth birthday, from my first interpretation of the data presented in Table 3.2.8 in section 3.2.1.2 (see the arguments given there). With this modification of Table 3.2.8, I note the following: If we calculate the average frequency of CP tokens produced for the 5 girls ranging in age from 9 to 14 years (201.2 tokens) and if we do the same for the 6 boys of the same age group (201.5 tokens), the difference in the production of CP tokens becomes a quantite negligeable, that is, insignificant. Consultants who are about 9 years of age and older show a balanced use of CP tokens with the three word classes elicited in the test. However, if we calculate the average number of CP types produced by the 5 girls and the 6 boys in age group II (9-14 years, note that consultant 24 is almost 9 years old), we find the following results (see Tables 3.2.8 and 3.2.124): For girls, the average number of CP types produced is 26.4; for boys, it is 29.8. Thus, boys in age group II produced more CP types than girls of the same age. As mentioned in sections 3.2.1.3-3.2.1.5 and as documented in Table 3.2.124, this gender difference holds for 4 of the 5 age groups tested. It is only with the consultants of age group IV that the average number of CP types produced by men does not surpass that for women. The gender-specific difference in the acquisition (and production) of CP types may be a result of the Trobrianders' education practices, which offer boys a much broader range of possibilities to experience their surrounding environment than they offer girls (see B. Senft 1985). Moreover, as my data show (see, e.g., G. Senft 1987d), men participate in public speeches much more often than women, although the women's social status in the Trobriand matrilineal society is rather high (see, e.g., Weiner 1976, 1988). Thus, it may be that parents and relatives especially promote their sons' or their young male relatives' rhetorical abilities and qualities. Here, we have to keep in mind that male adolescents usually develop a rather close relationship with their fathers and their male relatives. What should also be emphasized is that the Trobriand society, strictly stratified according to clan hierarchies,
The System of Classificatory Particles in Kilivila
183
offers social prestige to men who are born in lower-ranking clans if they prove to be good gardeners and/or good rhetoricians (see Malinowski 1935; Senft 1985e, 1987c, 1987d, 1991b; Weiner 1976, 1982, 1983, 1988). These social, extralinguistic facts may explain the gender differences we find in the acquisition and production of CP types—not only in children, but also in adolescents during and after puberty and in adults older than 36 years. Why we do not find this difference in the production of CP types in men and women in age group IV (21-35 years) is a question that will be discussed in section 3.3.3.3 below. Gender differences in the CP acquisition process that find their expression in preferences for the production of certain CP types will be discussed in the following section. It should be noted that the data documented in the overall corpus of Kilivila speech data confirm, at least in part, the results just presented, which were based on the analyses of the data elicited by the questionnaire (see sections 3.2.1.1-3.2.1.3; also 3.2.1.4 and 3.2.1.5).
3.3.2.4 What Is the Order in Which the Individual CPs Are Acquired? To answer this question, I will first present a table listing the CP types and their respective tokens produced by the consultants of the five age groups during the elicitation test. The table also lists the CP types that were not produced by the consultants in the elicitation test and the CP types and tokens produced in the overall corpus of Kilivila speech data that were not documented in the elicitation test data. I will now examine the individual CP types, following the order given in Tables 3.2.2, 3.2.9, and 3.2.16. In addition to these tables and Table 3.3.4, the tables presented in section 3.2.2 are relevant to the discussion of the order in which the individual CP types are acquired. The CP types to/te 'persons of male gender, human beings' and na (in its meaning 'persons of female gender' ( = na1) were produced by all consultants in age group I. As the tables show, only a few consultants produced more (or less) than the three expected tokens of these two CPs (these will be discussed in sections 3.3.3 and 3.3.4). On the basis of these data, I conclude that to/te and na1. are the first CPs that are acquired by Trobriand children. The CP kwe 'thing' was produced by 4 girls and 5 boys in age group I. Six of these 9 children show a rather high number of CP tokens produced during the elicitation test. Thus I conclude that children acquire this CP type with its broad range of accepted uses. The number of kwe tokens produced in the elicitation test rises quite dramatically with the consultants of age group IT; however, it decreases again with the consultants of age group III (these phenomena are discussed in sections 3.3.3 and 3.3.4). The CP ke 'wood, long objects, inanimates' was also produced by 4 girls and 5 boys in age group I. With this CP type, we observe phenomena that
184
Classificatory Particles in Kilivila
Table 3.3.4. CP types and tokens produced by all consultants No. of tokens produced by age group CP type
I
IV
III
II
beku 'stone blade'
1
V
1
bogi 'night'
2
8
2
bubulbobu 'cut across'
7
23 8
13
25
10 32
10 44 6
bubwa 'cut off' budalbodalbudu 'group' bukwa 'fruit cluster' bwa 'tree'
1 8
20
52
26
12
7
2
4
4
30
23 1
36
15 1
33 11
27 1
30 5
3
1
8
17
33
31
25 3
30
36
33
36
8
15
20
13
9 34
30
21
36
28 33
10
23
23
22 3
10
2 78
7 72 36
3 10
bwalita 'sea' deli 'group' doba 'grass skirt' duli 'cluster' duya/kaduya/kaduyo 'door'
3 1
gili 'row' giwi 'cut' guba 'bundle of taro' gudi 'child'
7
gula 'heap' gum 'bit' iga/yegila 'name' kabisi 'compartment'
10
kabulol 'point' kabulo2/kabulu 'sector' kada/keda 'road' kai 'stone blade' ka'i 'tooth' kala 'day' kapwa/kapo 'parcel' kasa 'row' kavi 'tool' ke 'tree, inanimates' kila 'hands of bananas' kova 'fire' kubila/kwabila 'land plot' kudul 'band of fibres' kudu2 'tooth' kumla 'earth oven' kwe 'thing' kweya/kwaya/keya 'limb'
17
2 81
3 13
27 30
93 35 28
5
18
6
11
32
46
204
499 2 36 12
380 4 36 26
3 36 804 20
8 36
17
19 305 3
553 60
35 26 6 37 412 1 35 29 3 9 36 643 69
36
30 3 11 39 3 424 35 27 6 6 35 527 63
The System of Classificatory Particles in Kilivila
185
Table 3.3.4. (continued) No. of tokens produced by age group CP type kwoila/kwela 'clay pot' kwoya/koya 'mountain'
I 5 1
liku 'compartment'
II
III
IV
V
26 4 11
14 11
26 7
3
22
19 2
25
lilo/lola 'walk'
1
3
19 1
luba 'bundle' luva 'wooden dishes'
1
9
21
25
21
1
3 1
8
6
5
6
2
8
1
34 39
39 67
39 54
36 51
3
12 1
33 6
24
3 31
19
34
36
9 35
35
11
39 3
64 18
67 11
68 21
pwanina/pona 'hole' sa 'nut bunch' si 'small bit'
17 1 5
24 3 5
36 3 5
31 3
32 6
sipu 'sheaf' sisi 'bough'
11 62
25
12
57
21 64
31 56
6 10
24
33
17
25 1
27 18
38
46
36 3
38 1
8
32
36 4 114
96 6
58 16
megwa 'magic' meila 'part of a song' mmwa 'conical bundle' nal 'female person' na2 'animals' nigwa 'hole' nina 'part of a song' notu/nutu 'kneaded, dot' nunu 'corner of a garden' peta/ta 'basket' pila 'part' po'ula 'grove'
siva 'time'
tuta 'time' utu 'scrap' uva 'span' va/vaya/vayo 'door' vakala 'belt' vili 'untwisted' vosi 'song' wela 'batch of fish' ya 'flexible, thin' yam 'day'
36 68 1
6
3
suya 'batch of fish' tam 'sprouting' tetu 'yams' to/te 'male, human'
3 12
1
7 1
22
3 88 4
9 1 1 26 6 14 43 11
8 1 25 10 18 54 4
24 2 7 66
186
Classificatory Particles in Kilivila
Table 3.3.4. (continued) No. of tokens produced by age group CP type yama/yuma 'hand' yulai/yule 'bundle of four things'
I
II 1
III
2
IV 2
6
3
yuva/yuwo 'shoal'
3
1
0 'basketful of yams'
1
1
CPs from the overall corpus of Kilivila speech data: bililo 'trip' dumia 'swamp' 1
4
kali 'paddle strike'
1
kauya 'creel'
4
lipu 'compartment/creel' na3 'stars, moon, planets'
5
V 7
4 4
na4 'carvings'
3
yeni 'handful'
1
CP produced but not assignable to individual consultants: sam 'ginger' CPs not produced in the elicitation test: lila 'bough', oyla 'string', siwa 'sea portions', vilo 'place'
are very similar to those mentioned for kwe. On the basis of these data, I conclude that the CPs kwe and ke come second in the CP acquisition process. The CPs kova 'fire', na in its meaning 'animals (= na2), ta 'basket', and ya 'flexible, thin' were produced by 3 girls and 3 boys in age group I. Although 1 consultant of age group II and 2 consultants in age group III did not produce the CP ya, the data justify the conclusion that 9-year-old children have acquired these CP types. Only with na2 and ya do we observe more tokens than expected in the elicitation test. For ya, it is again the consultants in age group II who show a rather dramatic rise in the number of tokens produced during the elicitation test. The CPs kumla 'earth oven' and pwanina/pona 'punctured, hole' were produced by 7 consultants in age group I. The data document that children in age group II have acquired the CP kumla. Five of the 6 boys in this age group have also acquired the CP pwanina/pona. Five consultants in age group I produced the CPs ka'i 'tooth', gudi 'child', and sisi 'bough'. The acquisition of the CP ka'i is completed in boys in age group II. All children in age group IT have acquired the CP sisi. With this CP, we also observe that the consultants in age groups II-V produced more tokens than expected in the elicitation test.
The System of Classificatory Particles in Kilivila
187
On the basis of these data, I conclude that the CPs kova, na2, ta, ya, kumla, pwanina/pona, ka'i, gudi, and sisi come third in the CP acquisition process. The acquisition of gudi, however, is earlier in boys than in girls. Two boys and 2 girls in age group I produced the CPs utu 'scrap' and iga 'name'. The acquisition of these two types is completed in children in age group II. With the CP utu, we observe again that more tokens than expected were produced by the consultants in age groups III-V. One boy and 1 girl in age group I produced tokens of the CP bwa 'tree'. Although we find 26 tokens of this CP produced by 6 consultants of age group II, it seems to play a secondary role in everyday speech, if we accept the token frequency as the tertium comparationis for such a judgment (this argument is explicitly described in section 3.3.3.1). Two boys and 1 girl produced tokens of the CP kada 'road' and si 'bit'. The CP si documented by only 5 tokens each produced by the consultants in the first three age groups; thus, it does not seem to play a large role in everyday speech. This is not the case with the CP kada, however, The acquisition of this CP is already completed in children in age group II. With this CP, we also observe that the consultants produced more tokens than expected. Three boys but only 1 girl produced tokens of the CP kasa 'row'. With this CP, we again observe that some consultants in age groups II-V produced more tokens than expected. The gender differences in the production of certain CPs is even more marked in the following cases: Buda 'group', duli 'cluster', kai 'stone blade', kweya 'limb', pila 'part, piece', and notu 'kneaded, dot' are produced only by boys in age group I. Boys in age group II are still ahead of girls in their acquisition of 4 of these 6 CPs, the exceptions being pila and buda. With the CPs pila, buda, and kweya, we again observe that more tokens than expected were produced by the consultants in age groups III-V. The CPs bubwa 'cut off', bukwa 'fruit cluster', kapwa 'parcel', kwoila 'clay pot', luba 'bundle', luva 'wooden dishes', and sa 'nut bunch' were produced by boys in age group I. Consultant 10 produced the only 3 tokens of the CP type bukwa that occurred during the elicitation test. In the overall corpus of Kilivila speech data, we find only 6 more tokens of this type, which were produced by the same consultant, 6-year-old Towesei. The CP bubwa does not seem to play a role in everyday speech, either. It was produced only once by consultant 12 in age group I. With age group III, we find 8 tokens of this type produced by only 3 consultants. As for the CP kapwa, we find 5 tokens produced by 2 boys in age group I. Three girls and 3 boys in age group II produced 18 tokens of this CP type. Only 3 girls in age group III produced 6 tokens of kapwa. Thus, this CP seems to play a secondary role in everyday speech. The same is true for the CP luva. We find only 18 tokens produced by the consultants. The CP sa was also documented by only 16 tokens. The CP luba was only produced once, by consultant 10. With the consultants in age group II, we find 9 tokens, and the consultants in age group III produced 21 tokens of this CP. For the CP kwoila/kwela, we find
188
Classificatory Particles in Kilivila
a few more tokens. Consultants 10 and 12 in age group I produced 5 tokens. Nine consultants in age group II produced 26 tokens of this CP. However, with the consultants in age group III, we find only 14 tokens. The CPs kaduyo 'door', kwoya/koya 'mountain', lilo/lola 'trip', va 'door', and yuma 'hand' were produced only by girls in age group I. However, as Table 3.3.4 shows, these CP types play a secondary role in everyday speech. For the sake of completeness, it should be noted that 6-year-old Towesei produced one token of the CP dumia 'swamp' in the overall corpus of Kilivila speech data. We can now summarize the results for CP acquisition in children ranging in age from 4 to 7 years, as follows. These children produced 37 CP types. The data show that the first CPs acquired are to/te 'male persons, human beings' and nal 'female persons'; the CPs kwe 'thing' and ke 'wood, long objects, inanimates' come second, and the CPs kova 'fire', na2 'animals', ta 'basket', ya 'flexible things', kumla 'earth oven', pwanina/pona 'punctured, hole' ka'i 'tooth', gudi 'child', and sisi 'bough' come third in these children's language acquisition process. In a fourth phase of this process, the CPs utu 'scrap', iga 'name', bwa 'tree', kada 'road', si 'bit', and kasa 'row' are acquired. With the other CPs acquired in this phase, we note marked gender differences. The data for the CPs gudi, kada, si, and kasa already pointed in this direction. In addition, we note that the CPs buda 'group', duli 'cluster', dumia 'swamp', kai 'stone blade', kweya 'limb', pila 'part', notu 'kneaded, dot,' bubwa 'cut off, bukwa 'fruit cluster', kapwa 'parcel', kwoila 'clay pot', luba 'bundle', luva 'wooden dishes', and sa 'nut bunch' are produced only by boys in this age group, whereas the CPs kaduyo 'door', kwoya/koya 'mountain', lilo/lola 'walk', va 'door', and yuma 'hand' are produced only by girls of this age group. As Table 3.2.214 shows, we find only 1 dubious and 8 unacceptable cases with these children's CP production. If we relate these 9 dubious or unacceptable responses to certain stimuli to the total number of tokens produced by the 12 children in this age group (860 tokens), we note that these dubious and unacceptable cases account for only 1.2%. Thus, we note that the children in age group I produced almost all their CPs in correct and acceptable contexts. Nevertheless, as shown in Table 3.2.5, the mean values with respect to cases of multiple classification were 14% for girls and 8% for boys in age group I; thus, linguistic insecurity exists in children of this age with respect to adequate and unequivocal CP production. With children ranging in age from 8 to 14 years, we find the following CP types first realized in the elicitation test: The CP giwi 'cut' was produced by 2 girls and 5 boys. Two girls and 4 boys produced tokens of the CP gum 'bit'. Two girls and 3 boys produced the CP liku 'compartment of a foodhouse'. The CP sipu 'sheaf was produced by 1 girl, but by 4 boys. The CPs bogi 'night' and kabisi 'compartment of a foodhouse' were produced by 2 girls and 2 boys. As Table 3.3.4 shows, the CP boti seems to play a secondary role in everyday speech production. Three boys, but only 1 girl, produced the CP kubila 'land plot'. The CP bubu 'cut across' was produced by 2 boys and 1 girl. The CPs deli 'group' and gula 'heap, group' are produced by 1 boy and by 1 girl.
The System of Classificatory Particles in Kilivila
189
Only boys produce the CPs tam 'sprouting', yam 'day', kabulo2 'sector', megwa 'magic', mmwa 'conical bundle', po'ula 'grove', and vili 'untwisted'. With these CPs, it should be noted that 3 boys produced the CP tarn and that 2 boys produced the CP yam. All other CPs were produced by 1 male consultant only. Moreover, Table 3.3.4 indicates that the CPs yam, kabulo2, megwa, mmwa, and po'ula play a rather small role in everyday speech production. Only girls in age group II produced the CPs suya 'batch of fish', ilia 'hands of bananas', kudu2 'tooth', nunu 'corner of a garden', and wela 'batch of fish'. The CP suya was produced by 2 girls. All other CPs are produced by 1 female consultant only. Moreover, as Table 3.3.4 indicates, these CPs, with the exception perhaps of suya seem to play a secondary role in everyday speech production. It should be noted that the CP kudu2 is synonymous with the CP ka'i; the acquisition of this CP is completed in boys in age group II. In addition to these CP types, the consultants in age group II produced all the CP types produced by the children in age group I except the CP types bubwa 'cut off', bukwa 'fruit cluster', dumia 'swamp', and lilo 'walk', which seem to play a secondary role in everyday speech production. Summarizing the results for CP acquisition in children ranging in age from 8 to 14 years, we note that these children produced 55 CP types; they acquired 22 new CP types. The production of these CP types shows rather pronounced gender differences in these children's CP acquisition process. Boys produced 17 of the 22 newly acquired CPs; they acquired 13 of them earlier and faster than girls in this age group. This result is especially marked for the CPs giwi 'cut', gum 'bit', sipu 'sheaf, liku 'compartment of a foodhouse', kubila 'land plot', and tam 'sprouting'. Girls produced 15 of the 22 newly acquired CP types. However, as we can infer from the data presented, they were just starting to acquire these CP types. As Table 3.2.214 shows, there were 10 dubious and 22 unacceptable cases of CP production in this age group. If we relate these 32 cases to the overall number of CP tokens produced by the 12 consultants in this age group (2,341 tokens) we note that these dubious and unacceptable cases account for only 1.4% of the CPs produced. Thus, we note that the children in age group II produced almost all acquired CPs in correct and acceptable contexts. Nevertheless, Table 3.2.12 shows that mean values with respect to cases of multiple classification were 32% for girls and 28% for boys in age group II; thus, linguistic insecurity exists in children of this age with respect to adequate and unequivocal CP production. With the adolescents ranging in age from 15 to 20 years, we note the following results: The consultants in age group III produced 64 CP types. Three CP type variants and 10 CP types proper were produced in the elicitation test for the first time by the consultants of this age group: The variant duya of the CP duya/kaduya/kaduyo 'door, entrance' and the variant vaya of the CP va/vaya/vayo 'door' were produced by 1 male consultant in each case. Two female and 2 male consultants produced the variant yegila
190
Classificatory Particles in Kilivila
of the CP iga/yegila 'name'. The other variants of these 3 CP types, however, can already be found in the language production of the consultants in age group I. The CPs bwalita 'sea' and tuta 'time' were each produced by 1 male and 1 female consultant. Two female and 2 male consultants produced the CP vosi 'song'. The CPs uva 'span', yulai 'bundle of four things', gili 'row', and 0 'baskets of yams' were only produced by male consultants. Only 1 female consultant in each case produced the CPs beku 'stone blade', vakala 'belt', and yuva 'shoal'. Thus, we still find gender-differences in this age group in the production and acquisition of CP types. However, as Table 3.3.4 shows, all these CP types and CP type variants seem to play a secondary role in everyday speech production. As Table 3.2.214 shows, there were 5 dubious and 14 unacceptable cases of these consultants' CP production. If we relate these 19 cases to the overall number of CP tokens produced by the 12 consultants in this age group (2,450 tokens), we note that these dubious or unacceptable cases account for only 0.8% of all CPs produced. Thus, the adolescents in age group III produced almost all their CPs in correct and acceptable contexts. Nevertheless, Table 3.2.19 shows that mean values with respect to cases of multiple classification were 21% for female and 12% for male consultants in age group III; thus, linguistic insecurity exists in consultants of this age with respect to adequate and unequivocal CP production. For the sake of completeness, I want to note that the CPs kudul 'band of fibers', kauya 'creel', doba 'grass skirt', lipu 'compartment of a creel', and tetu 'yams' were only found in the language production of the consultants in age group IV. Moreover, the CPs kabulol 'protuberances', na3 'moon', and na4 'carvings in human likeness', kavi 'tool', kala 'day', meila 'part of a song', nigwa 'hole', nina 'part of a song', siva 'time', bililo 'trip', kali 'paddle strike', and yeni 'a handful of something' were only found in the language production of consultants in age group V. It should be noted here, however, that the CP kudul 'band of fibers' was produced by 1 woman in age group IV but by 2 men in age group V. The CP kabulol 'protuberances' was produced only by 1 woman. All the other CP types just mentioned were produced by only 1 man in age groups IV and V. Thus, we again find marked gender-differences in the CP production of these consultants. However, as Table 3.3.4 documents, all these CPs seem to play a secondary role in everyday speech production. I will discuss these phenomena, together with the problem of dubious and unacceptable cases of CP production, in section 3.3.3. I can now summarize the answer to the question concerning the order in which the individual CPs are acquired by listing the CPs in Table 3.3.5 according to this order, expressed in six phases. Although I elicited the production of the CPs lila 'bough, branch, leaf, oyla 'string', siwa 'sea portions', and vilo 'place' in the informal pretests, these CPs were not produced during the elicitation test proper; moreover, these
The System of Classificatory Particles in Kilivila Table 3.3.5. The order of CP type acquisition Age group/phase
CP type 'gloss(es)' (comment)
I/1
tolte 'male persons, human beings' nal 'female persons'
I/2
ke 'wood, long objects, inanimates' kwe 'thing'
I/3
kova 'fire' na2 'animals' ta 'basket' (see: III/6} ya 'flexible, thin' kumla 'earth oven' pwanina/pona 'punctured, hole' ka'i 'tooth' gudi 'child' (acquired earlier by boys) sisi 'bough'
I/4
utu 'scrap' iga 'name' (see: III/6) bwa 'tree' kada 'road' (acquired earlier by boys) si 'bit' (acquired earlier by boys) kasa 'row' (acquired earlier by. boys) buda 'group' (acquired by boys only) duli 'cluster' (produced by 1 boy only) dumia 'swamp' (produced by 1 boy only) kai 'stone blade' (produced by 1 boy only) kweya 'limb' (produced by 1 boy only) pila 'part' (acquired by boys only) notu 'kneaded' (acquired by boys only) bubwa 'cut off (produced by 1 boy only) bukwa 'cluster' (produced by 1 boy only) kapwa 'parcel' (acquired by boys only) kwoila 'clay pot' (acquired by boys only) luba 'bundle' (produced by 1 boy only) luva 'dishes' (produced by 1 boy only) sa 'nut bunch' (produced by 1 boy only) kaduyo 'door' (see III/6; produced by 1 girl only) kwoya/koya 'mountain' (produced by 1 girl only) lilo/lola 'walk' (produced by 1 girl only) va 'door' (see III/6; produced by 1 girl only) yuma 'hand' (produced by 1 girl only)
II/5
giwi 'cut' (acquired earlier by boys) gum 'bit' (acquired earlier by boys) sipu 'sheaf (acquired earlier by boys) liku 'compartment' (acquired earlier by boys) kubila 'land plot' (acquired earlier by boys) bubu 'cut across' (acquired earlier by boys) bogi 'night' kabisi 'compartment of a foodhouse'
191
192
Classificatory Particles in Kilivila
Table 3.3.5. (continued) Age group/phase
CP type 'gloss(es)' (comment) deli 'group' gula 'heap' tam 'sprouting' (acquired by boys only) yam 'day' (acquired by boys only) kabulo2 'sector' (produced by 1 boy only) megwa 'magic' (produced by 1 boy only) mmwa 'bundle' (produced by 1 boy only) po'ula 'grove' (produced by 1 boy only) vili 'untwisted' (produced by 1 boy only) suya 'batch/fish' (acquired by girls only) kila 'hands of bananas' (produced by 1 girl only) kudu2 'tooth' (produced by 1 girl only) nunu 'corner of a garden' (produced by 1 girl only) wela 'batch/fish' (produced by 1 girl only)
HI/6
duya 'door' (see I/4; produced by 1 man only) vaya 'door' (see I/4; produced by 1 man only) yegila 'name' (see I/4) peta 'basket' (see I/3) bwalita 'sea' tuta 'time' vosi 'song' uva 'span' (acquired by men only) yulai 'bundle of four things' (acquired by men only) gili 'row' (produced by 1 man only) 0 'basketful of yams' (produced by 1 man only) beku 'stone blade' (produced by 1 woman only) vakala 'belt' (produced by 1 woman only) yuva 'shoal' (produced by 1 woman only)
Additional CP types produced by consultants in age groups IV and V: IV
kudul 'band of fibres' (in IV produced by 1 woman; in V produced by 1 woman and 2 men) doba 'skirt' (produced by 1 woman only) kauya 'creel' (in IV and in V produced by 1 man only) lipu 'compartment of a creel' (produced by 1 man only) tetu 'yams' (produced by 1 man only)
V
guba 'bundle/taro' (produced by 1 woman only) kabulol 'protuberances' (produced by 1 woman only) na3 'moon' (produced by 1 man only) na4 'carvings in human likeness' (produced by 1 man only) kavi 'tool' (produced by 1 man only) kala 'day' (produced by 1 man only) meila 'part/song' (produced by 1 man only) nigwa 'hole' (produced by 1 man only) nina 'part/song' (produced by 1 man only) siva' 'time' (produced by 1 man only) bililo 'trip' (produced by 1 man only) kali 'paddle strike' (produced by 1 man only) yeni 'handful of something' (produced by 1 man only)
The System of Classificatory Particles in Kilivila
193
CPs were not produced in my overall corpus of Kilivila speech data. Thus, I cannot provide any information with respect to the acquisition of these CP types. The CP sam 'ginger' is documented in Senft and Senft (1986:142) in data on Trobriand string figures and the verses that accompany these figures. In the present study, the CP sam was produced by a 6-year-old boy and a girl of the same age. It is not possible to draw any inferences about the order in which the CP types are acquired from this small sample for the CP sam. In summary, the following main trends can be noted regarding the order in which CPs are acquired: 1. CPs that distinguish sex in human beings are acquired first. 2. CPs that take over the function of unmarked forms for inanimates are acquired next. The overgeneralization of these CPs is documented in our data. 3. The CPs that are acquired next denote concrete, specific objects and certain features of objects. 4. The fourth phase of CP acquisition adds CPs that are characteristic of the third phase, as well as CPs that denote parts, groups, clusters, and specific features of certain objects. 5. The characteristics of the fourth phase of CP acquisition become even more marked in the fifth phase. In addition, we find CPs that denote abstract concepts. 6. The final phase of CP acquisition increases the inventory of CPs characteristic of the fourth and fifth phases. 7. Boys not only acquire more CP types than girls, but also acquire these types earlier than girls (see section 3.3.2.3). 8. The majority of CPs that are produced by adults who have finished their first language acquisition can be characterized as technical terms of different terminologies. We may regard all these CPs as evidence for the continuing acquisition process with respect to the vocabulary of a language. At this stage of the analysis, I must postpone any attempt to answer the question of why the CPs are acquired in the order just presented. I will try to answer this question after having defined the semantic domains that are established and covered by the individual CP types (see section 3.3.4.3). 3.3.2.5 Are There Any Parallels Between Kilivila CP Acquisition and CP Acquisition Data for Other Classifier Languages? As mentioned in section 3.1, there are very few studies that deal with this question in other classifier languages (if we leave aside literature on sign languages such as American Sign Language). Nevertheless, if we compare the results presented by Gandour et al. (1984), Erbaugh (1986), Carpenter
194
Classificatory Particles in Kilivila
(1991), Matsumoto (1985), Sanches (1977), Chin (1989), Luke and Harrison (1986), and all the remarks on this topic in Allan (1977:295) quoting Burling (1965:261) and Haas (1942:201) with the answers to the questions raised in the previous sections, some interesting parallels emerge. Burling's observation of an overgeneralizing use of the Burmese classifier khu in children "in situations in which an adult would probably use a more precise classifier" (Burling 1965:261) parallels my observation of a rather dramatic rise of tokens for a number of CP types, but especially for the CPs kwe 'thing, unmarked form for inanimates' and ke 'wood, long objects, unmarked form for inanimates') in the language production of children in age groups II and III. From these data, I infer the Trobriand children produce at least these two CPs in situations in which an adult, or even an adolescent, would use a more precise classifier. Moreover, the observed "tendency to use the general classifier instead of the specific classifier" (Gandour et al. 1984:465, 468), the overgeneralization and "semantic overextension" (Carpenter 1991:105) of certain CPs in Thai and Japanese children (Carpenter 1991; Matsumoto 1985; Sanches 1977:58-59), and the "tendency for overgeneralization" of certain noun classes documented in Bantu language acquisition data (Demuth et al. 1986:463-465), as well as the fact that Chinese children "overwhelmingly preferred the general ge classifier" (Erbaugh 1986:413), are in agreement with my observations, especially with respect to the Trobriand children's production of the two CP types mentioned above, namely kwe and ke. However, with the acquisition of noun classes in Sesotho, a Bantu language, there are no cases of such overgeneralization (see Demuth 1988:310, 316; 1992:591, 594, 596-597). In connection with Chinese (Mandarin) noun classifiers, Erbaugh (1986:412) emphasizes "overheard adult narratives are a more important source of classifier input than speech directed to the child," and that the "majority of classifiers appeared in formal storytelling and songs" in adult-child conversations. Moreover, she notes that "children used special classifiers in exactly the same discourse settings as the adults" (1986:425) and points to "discourse sensitivity" (1986:426) in the Chinese children's use of classifiers. Erbaugh's observations are similar to my observations with respect to the specific use of CPs with demonstrative pronouns, adjectives, and numerals in Trobriand children ranging in age from 4 to 14 years. In section 3.3.2.2, I especially emphasized the role of the text category "fairy tale" for the acquisition process of different CP types used in the word formation of demonstrative pronouns. I also mentioned that the CP sam 'ginger', for example, was only used in connection with a string figure accompanying verse. This discourse or text category sensitivity observed with the CP production of Trobriand and Chinese children may also support the strategy mentioned by Haas (1942:201) in connection with classifiers in Thai: "...it is desirable to memorize the classifier to be used with a noun at the same time that one hears the noun...". The rather frequent use of certain CPs in fairy tales, songs, and verses, where the word classes that use CPs in their word formation
The System of Classificatory Particles in Kilivila
195
are immediately followed by the noun the CP refers to, may indicate that Trobriand adults as well as the children citing these verses, singing these songs, or telling these fairy tales apply just the strategy described by Haas to foster the acquisition process of CPs, especially with demonstrative pronouns. Moreover, the Trobianders seem to use this strategy to foster the acquisition of more special or specific CP types. As to these more special classifiers, Erbaugh (1986:415) records that with Chinese children "special classifiers developed slowly and stayed rare between the ages of 1.10 and 3.10." She notes more specifically (1986:417) that the Chinese children investigated "almost never used special classifiers before the age of 2.6." If I compare my observations with Erbaugh's results, I have to note that in Trobriand children these special classifiers develop even more slowly than in Chinese children (cf. also Carpenter 1991:103, Matsumoto 1985). However, the observations of Gandour et al. that in Thai children "...the rate of acquisition of ... classifiers was slow. ...10-year-old Thai children have still not mastered the classifier system" (Gandour et al. 1984:460) and Chin's (1989:92) findings on the acquisition of classifiers in Hokkien agree with my results. My data show that we can with good reason assume that the core of the Kilivila CP system is acquired at the end of puberty. The fact that "repeaters and general classifiers are ... easier ... to learn" (Gandour et al. 1984:463) and that "animate classifiers... are acquired earlier than classifiers in other categories" (Gandour et al. 1984:459) also holds for Thai and Hokkien (Chin 1989:123), Mandarin, and Cantonese (Luke and Harrison 1986) children, as well as for Trobriand children. With these few observations, I must conclude the discussion of results reported in other studies, which agree at least in part with my findings on the Kilivila CP acquisition process. 3.3.2.6 Excursus: Are There Parallels Between the Kilivila Data and Acquisition Data With Respect to Demonstrative Pronouns, Adjectives, and Numerals for such Indo-European Languages as English and German? I want to conclude section 3.3.2 with a few general and rather anecdotal remarks on the question of whether we can find any parallels between the results reported so far and acquisition data with respect to demonstrative pronouns, adjectives, and numerals for Indo-European languages like English and German. This question is so complex that this excursus cannot claim to provide anything more than rather cursory information on the topic. The literature cited in the following paragraphs was chosen more or less at random, just to indicate the course the argument will take. Thus, I want to emphasize that I am completely aware that my proceeding here does not adhere to the usual psycholinguistic standards. Nevertheless, I hope that the following anecdotal remarks may serve as an incentive to linguists to overcome the disproportion between research on Indo-European languages, on the one hand, and other so-called exotic languages, on the other hand.
196
Classificatory Particles in Kilivila
In their classic monograph, Stern and Stern (1928:280) emphasize the following: "Die Zahlworter treten dem auBeren Anschein nach relativ fruh auf; doch verbindet sich erst spat mit ihnen wirklicher Zahlensinn.... So 'zahlte' unsere Eva 2;0 ziemlich richtig bis zwolf, naturlich ohne jeden Zahlensinn". As to the acquisition of adjectives, they state: "Ferner aber zeigen die Listen von SCUPIN, NICE und DESCOEUDRES vollig uberinstimmend, daB die Adjektiva in der gleichen Zeit noch nicht Stabilitat gewonnen haben, sondern um 3;9 herum und noch jahrelang spater einen prozentualen Aufstieg zeigen, der ihren relativen Anteil um die Halfte gegenuber dem Stande um 2;0 erhoht" (Stern and Stern 1928:245). With the exception of personal pronouns, Stern and Stern (1928:279) remark: "Die ubrigen Pronomina scheinen uns keine Entwicklungsphanomene von allgemeinerem psychologischen Interesse zu bieten." Thus, with respect to demonstrative pronouns, we find only anecdotal remaks about their usage in the researchers' own children at the ages of 2 years 4 months (1928:100), 2 years 6 months (1928:54), and 3 years 2 months (1928:66). However, Bohme, who studied children between 4 and 5 years 6 months of age, emphasizes the "relatively late acquisition" of pronouns that are "variable in their referential function" (Bohme 1983:191). If we compare these statements with our Kilivila data, we can note that in Indo-European languages like German, numerals and adjectives are also acquired before demonstrative pronouns. Having noted this parallel between German and Kilivila acquisition data, we must emphasize that the process of the acquisition of the three word classes starts later and develops more slowly in Kilivila than in German. If we compare the Kilivila data with the results on the order in which number and countability distinctions are acquired by English-speaking children reported by Mufwene (1984:207-208), we again find certain parallels; albeit with the same time shift in Kilivila as emphasized previously. Tversky (1986:72), on the basis of acquisition data from studies on English-speaking children, states: The perceptual features children use as a basis for categorization in preference to the functional features characterizing! superordinates are also the features common in many classifiers. ... It has been proposed . . . that children prefer to group different basic level objects by perceptual features than by functional features because unlike functional features, perceptual features are readily observable; they are also the features successfully used in basic level categorization. According to this argument, children shift to functionally-based groupings because the latter are more informative than perceptually-based groupings. My Kilivila data support this proposition to a certain degree. I will use it as a starting point for trying to answer the question of why the individual CPs are acquired in the order described in section 3.3.2.4 (see section 3.3.4.3).
The System of Classificatory Particles in Kilivila
197
Finally, I would like to emphasize that the Kilivila data also support, at least to a certain degree, the developmental sequence proposed by the "common assumption in philosophy... and psychology" that Tyler (1969:19) quotes with so much reservation and suspicion. With respect to Kilivila CP acquisition, it seems to be the case that the "child first acquires concepts for concrete objects and gradually expands its semantic domain by extension and generalization to include more abstract conceptions" (Tyler 1969:19). However, we will discuss this problem once more in section 3.3.4.3.
3.3.3 What Is the Realization of the Individual CP Types in Actual Speech? To answer this complex question, I will first look at the rate of frequency with which the individual CP types were produced by the consultants during the elicitation test and in my recorded texts. This analysis will result in a tripartite classification of the CP types, the classes being characterized as follows: 1. CPs that form the core of the Kilivila classifiers produced in actual speech. 2. CPs that play a secondary role in actual speech. 3. CPs that play a secondary role in everyday speech because they belong to a special technical language variety or because they are extremely rare or even "almost obsolete" (Malinowski 1920:44). I will then look at the individual CP types and check Tables 3.2.126-3.2.212 with respect to the question of which CP type or types were actually produced versus which were expected during the elicitation test. This analysis is a first step leading to a definition of the semantic domains established by the system of Kilivila CPs. Thus, this analysis is a prerequisite to the analyses that will be presented in section 3.3.4. However, this analysis will also enable us to answer the question of whether we can observe any language change in progress that affects this complex CP system. Attempts to answer this question involve discussion of the following problems: 1. Which CP types are affected by such a language change? 2. Why are they affected? 3. What about the use of certain CP types that are produced as stylistic variables or variants? 4. Is the usage of certain CP types determined by the age and/or status and/or gender of the speaker? The section ends with some speculations about the consequences any observed changes might have for Kilivila grammar as a whole.
198
Classificatory Particles in Kilivila
3.3.3.1 Can a Classification of the CP System Be Devised? One of the most obvious and straightforward observations that can be made about the data presented is the enormous differences between the number of tokens produced for the individual CP types. In section 3.3.2, I argued that, on the basis of token frequency, some CP types or CP type variants seem to play a secondary role in everyday speech production. In the elicitation test, I asked 60 consultants to produce one token of each type with three different word classes. Thus, the maximum of tokens to be elicited for one CP type was, at least in theory, 180 tokens per type (= 60 consultants x 1 token each for 3 word classes). If I elicited more than 180 tokens per type in all consultants, it would mean that one or more consultant(s) produced one or more token(s) of the respective CP type as a response to one or more stimulus/stimuli presented to elicit a different CP type. If I cannot exclude these CP usages on the basis of Table 3.2.214 as dubious or unacceptable cases of CP production, I infer that such CP usage documents the semantic scope of the CP type. If the number of tokens elicited surpasses the number 180 to a high degree, I infer that this CP type is a kind of "general" classifier, a classifier used in an overgeneralized way, or a classifier with a rather broad (though not general) semantic scope. If the number of tokens elicited fell well below 180, I inferred that this CP type does not play an important role in everyday speech production. If the number of tokens elicited fell below 180 to an extreme degree, I inferred that this CP type plays a secondary role in actual speech because it belongs to a special technical language variety, is only very rarely used, or is almost obsolete. I want to emphasize once again that I have not used statistical analyses because the data show only a nominal scale standard. Thus, in interpreting the data I am confronted with problems of the following kind: In Table 3.2.125, for example, I count 42 tokens for the CP we/a; for the CP uva I count 31 tokens. Is the difference of 11 tokens large enough to justify a different classification of these two CP types? This raises the classic procedural question: "What can we go by?" The arguments that follow are based on the evidence of the data presented in the tables. Thus, I must leave it to the critical reader's discretion whether or not he or she agrees with my interpretations. I have demonstrated elsewhere (G. Senft 1982) that this pragmatic proceeding leads to presentable results. Keeping these considerations in mind, we can now start to answer the question raised at the beginning of this section. Table 3.2.125 lists 84 CP types or CP type variants, ordering them according to the number of tokens produced for each. For the sake of a more complete description, this table also presents percentages based on the actual number of tokens produced. I interpret these data in the following way: 1. The exceptional status of the two CPs kwe 'thing, inanimates' and ke 'wood, long objects, inanimates' is obvious. With their function as
The System of Classificatory Particles in Kilivila
199
unmarked forms for inanimates, their semantic scope seems to be that of so-called general classifiers. 2. The CPs kada 'road', utu 'scrap', na1 'female person', na2 'animals' ya 'flexible', sisi 'bough', pila 'part', kweya 'limb', and to/te 'male persons, human beings' play an important role in the consultants' actual speech. 3. If we compare the tokens and percentages noted for the CP types/variants na1 'female persons' and kasa 'row', on the one hand, and ka'i 'tooth' and duli 'c'uster', on the other hand, in Table 3.2.125, we note that the CP types/variants kasa 'row', kumla 'earth oven', kova 'fire, buda/boda 'group', ta 'basket', gudi 'child', iga 'name', pwanina/pona 'hole', kai 'stone blade', and ka'i 'tooth' are an essential part of the CP inventory of actual speech. 4. If we compare the tokens and percentages noted for the CP types gum 'bit' and sipu 'sheaf, on the one hand, and kabisi 'compartment', on the other hand, in Table 3.2.125, and if we concede that CPs whose expected production totals almost 50% in actual speech can be incorporated into the group of CP types that form the core of the Kilivila CP system, we have to mention the CP types duli 'cluster', giwi 'cut', notu, nutu 'kneaded, dot', kwoila/kwela 'clay pot', kubila 'land plot', deli 'group', suya 'batch of fish', gum 'bit', and sipu 'sheaf here, too. Thus, on the basis of the data presented in Table 3.2.125, we can argue that the following 29 CP types/variants constitute the core of the Kilivila CP system realized in actual speech: kwe, ke, na1 + 2, kada, utu, ya, sisi, pila, kweya, to/te, kasa, kumla, kova, buda/boda, ta, gudi, iga, pwanina/pona, kai, ka'i, duli, giwi, notu/nutu, kwoila/kwela, kubila, deli, suya, gum, and sipu. Although the CP types kabisi 'compartment', luba 'bundle/taro', vili 'untwisted', liku 'compartment', tam 'sprouting', bubu/bobu 'cut across', bwa 'tree', gula 'heap', po'ula 'grove', kapwa/kapo 'parcel', and wela 'batch of fish' are realized more frequently than the CP types/variants uva 'span', bubwa 'cut off, kudu2 'tooth', kwoya/koya 'mountain', mmwa 'conical bundle', nunu 'corner/garden', kabulo2 'sector', yam 'day', vosi 'song', and luva 'wooden dishes', I think it is justifiable to group these 21 CP types/variants together and to note that they seem to play a secondary role in actual speech. If we keep in mind that duya and kaduyo 'door, entrance' are variants of one CP type, we can include this CP type in this second group, too. Thus, this group contains 22 CP types. Although the line I will draw between these 22 CP types/variants and the secondary CP types/variants may seem to be unclear, I think it is justifiable to group together those CPs with a realization of less than 10%. Thus, I conclude that the following CP types/variants play a secondary role in everyday speech because they belong either to a special technical language variety (e.g., gili 'row', megwa 'magic', kudul 'band of fibers', meila 'part/song', nina 'part/song' beku 'stone blade', vakala 'belt', vakala 'belt', and 0 'basketful of yams') or because they are extremely rare or almost obsolete (e.g., sa 'nut
200
Classificatory Particles in Kilivila
bunch', si 'bit', va/vaya/vayo 'door', bogi 'night', yama/yuma 'hand', yulai/yule 'bundle of four things', tuta 'time', kila 'hands of bananas', lilo/lola 'walk', peta 'basket', yegila 'name', bwalita 'sea', yuva 'shoal', bukwa 'fruit cluster', guba 'bundle/taro', kabulol 'point', kala 'time', kavi 'tool', siva 'time', doba 'grass skirt', nigwa 'hole', and tetu 'yams'). Considering the CP that occur in my recorded and transcribed texts (presented in Table 3.2.218), I can incorporate the CP types/variants kauya 'creel', lipu 'compartment/creel', bililo 'trip', dumia 'swamp', kali 'paddle strike', na3 'moon', na4 'carvings in human likeness', sam 'ginger', and yeni 'a handful of something' into the last group. Moreover, the CP types lila 'bough', oyla 'string', siwa 'sea portions', and vilo 'place' that were not produced during the elicitation test and were not documented in the recorded texts also belong to this third group. Table 3.3.6 summarizes the results of this section. In his article "Classificatory Particles in the Language of Kiriwina," Malinowski classified the formatives he presented according to their "degree of obsoleteness" (Malinowski 1920:55-57). I will conclude this section by comparing Malinowski's results and remarks with my findings. First, it is evident to anyone familiar with Malinowski's article that my spelling of some of the CPs is different from his. These differences document phonological changes within the last 65 years or so, although I do not exclude the possibility that some of these differences may be a result of language varieties. Looking at some individual CP types, I note the following: My data on the CP sisi 'bough' imply, contrary to Malinowski, that this CP is capable of semantic extension. Table 3.2.190 shows that sisi was quite often produced when the CP ya 'thin, flexible' was expected. The CP lila 'bough, branch, leaf is replaced now by the CP ke 'wood, inanimates, long objects', as Malinowski assumed it would be (see Table 3.2.159); however, my pretests showed that my consultants still know this CP. The CP siwa 'sea portions' is also no longer produced by my consultants, as Malinowski predicted; it has been replaced by kwe 'thing'. The CP siva, as well as the CP kala, however, are no longer "perfectly vital" (Malinowski 1920:56), as Malinowski reported. This observation also holds for the CPs sa 'nut bunch', si 'bit', and 0 'basketful of yams' (but see the discussion of this CP type in 3.3.4.2). Thus, I note the following differences between Malinowski's and my observations on the Kilivila CP system, leaving aside the phonological changes mentioned: 1. Some formatives Malinowski mentioned are rarely or never used by my consultants. 2. With some formatives, I found differences between Malinowski's definition of the CP and my consultants' actual production of them. However, it should be emphasized that with the majority of the CP types Malinowski mentioned, my findings agree more or less with his results,
The System of Classificatory Particles in Kilivila
201
Table 3.3.6. Classification of the CP system in Kilivila
The core of CP types/variantst
CP types/variants that play a secondary role in actual speech
kwe ke
kabisi luba
CP types/variants that belong to a technical language variety or are extremely rare or almost obsolete
sa va/vaya/vayo
si bogi
*nal + 2
vili
gili
megwa
kada
liku
yama/yuma
*kudul
utu
tam
yulai/yule
tuta
ya sisi
bubu/bobu
kila
lilo/lola
bwa
*peta
*yegila
pila
gula
bwalita
yuva
kweya
po'ula
bukwa
guba
to/te
kapwa/kapo
*kabulol
kala
kasa
wela
kavi
meila siva
kumla
uva
nina
kova
bubwa
beku
vakala
buda/boda
*kudu2
0
doba
*ta
kwoya/koya
nigwa
tetu
gudi
mmwa
kauya
lipu dumia
*iga
nunu
bililo
pwanina/pona
*kabulo2
kali
*na3 + 4
kai ka'i duli
yam vosi luva
sam lila siwa
yeni
giwi
kaduya/duya
oyla vilo
nutu/notu kwoila/kwela kubila deli suya
gum sipu 29 CP types/variants
22/CP types/variants
42 CP types/variants
Variants are marked with an asterisk.
remarks, and predictions, at least with respect to their "degree of obsoleteness". This observation once more highlights this great scholar's exceptional scientific qualities.
202
Classificatory Particles in Kilivila
3.3.3.2 How Do the Consultants Realize the Individual CP Types and Can We Observe Any Processes of Language Change in Progress? To answer these questions, I will first look at the individual CP types presented in Tables 3.2.37-3.2.123 and especially in Tables 3.2.126-3.2.212 and check which types were actually produced when I expected certain CPs during the elicitation test. This procedure will serve as the basis for a discussion of the observable processes of language change in progress, namely: 1. Which CP types are affected by such a language change? 2. Why are they affected? 3. What about the use of certain CP types that are produced as stylistic variables or variants? 4. Is the usage of certain CP types determined by the age, and/or status, and/or gender of the speaker? To provide answers to all these questions, of course, I must take into account all the relevant tables presented in section 3.2. In the following paragraphs, I will discuss these questions for the individual CP types. Table 3.2.126 showed that only 2 tokens of the CP type beku 'stone blade' were produced by 2 consultants instead of the expected CP kai 'stone blade'. If we consider Tables 3.2.37 and 3.2.124 and appendix A, we realize that the number of types the 2 consultants Itakeda (consultant 27) and Bwetadou (consultant 45) produced is clearly above average compared with the speech behavior of other consultants in that age group. Itakeda is a member of the highest-ranking clan, Malasi; Bwetadou is heir to his father's (!) weather magic, and thus also has a relatively high social status, although he is a member of the lowest-ranking clan. Thus, on the basis of these considerations, we may tentatively assume that the use of this CP may serve the speakers as a means to mark their societal status. They produce an uncommon "repeater", which is extremely rare, as a stylistic variant together with the respective tokens of the actually expected and elicited CP type kai (see Table 3.2.63). The CP type bililo 'trip' was produced 4 times by 41-year-old Tosulala describing the construction of a canoe (see Tables 3.2.216, 3.2.217, and appendix B). Tosulala is a member of the highest-ranking clan. I assume that the CP bililo is almost obsolete and that consultant V17 produced it in his thorough and careful description as a stylistic variant. As Tables 3.2.38 and 3.2.127 show, only 12 tokens of the CP type bogi 'night' were realized by 7 consultants, 4 of them in age group III. The consultants 26, 31, 34, and 38 are members of the highest-ranking clan. Consultant 45 is heir to the weather magician of Tauwema village, and consultant 21 is the chief's grandson. Only consultant 27 does not have any special status within the village (see appendix A). Thus, 1 infer that this CP is produced by villagers of status as a sociolinguistic variable, as it is defined
The System of Classificatory Particles in Kilivila
203
by Labov (1972b:237).4 The repeater is obviously replaced by the CP type kwe 'thing' (see Table 3.2.166). The frame of reference of the CP bogi is distinct, of course. Tables 3.2.40 and 3.2.129 confirm our finding that the CP bubu/bobu 'block cut across/cut off plays a secondary role in actual speech production. As the comments to Table 3.2.129 indicate, this CP type seems to have become a bit problematic with respect to a clearcut and adequate usage in connection with nominal references. We find 4 unacceptable reactions of the consultants to the stimulus presented to elicit this CP type; moreover, the consultants produced the CP types kwe 'thing', ke 'wood, long objects, inanimates', ya 'flexible', bwa 'tree', and bubwa '(parts) cut off when bubu/bobu was expected. It seems that the CP kwe, but also the CP ke, have started to supersede this CP type. As Table 3.2.166 shows, the use of kwe instead of bubu/bobu can be observed rather markedly in consultants in age groups II and IV. As Tables 3.2.41 and 3.2.130 document, the CP bubwa 'parts cut off plays a secondary role in actual speech. Table 3.2.130 shows that the frame of reference of this CP type is not definite, but rather is quite fuzzy. My consultants produced the CP bubwa when I expected the CPs utu 'scrap', bubu 'cut across', kabulo1 'point', and vili 'untwisted', although these last two cases of bubwa are rather dubious, if not unacceptable (see Table 3.2.214). The CP type ke 'wood, long objects, inanimates' seems to have superseded the CP bubwa. If we compare Table 3.2.41 with the data in appendix A, we find the following: 8 of the 14 consultants producing the CP bubwa belong to the highest-ranking Malasi clan (consultants 25, 31, 38, 41, 47, 49, 56, and 60); Toybokwatauya (consultant 43) and Sogeya (consultant 50) are chief Kilagola's children. Thus, I note that 10 of the 14 consultants that produced the CP bubwa are of high status in Tauwema society. Although we find the dubious cases of bubwa production with two of these consultants (consultants 25 and 41), I nevertheless infer that the CP bubwa is produced by villagers of status as a sociolinguistic variable. This finding is confirmed by Tables 3.2.216, 3.2.217, and appendix B. The Tables 3.2.42 and 3.2.131 show that the CP type buda/boda/budu 'group, team' not only belongs to the core group of Kilivila CP types (see also Tables 3.2.216, 3.2.217, and 3.3.6), but also seems to take over the referential function of the CP yuva 'shoal'. However, this finding was expected 4
Labov (1972:237) states, "We may define a sociolinguistic variable as one which is correlated with some nonlinguistic variable of the social context: of the speaker, the addressee, the audience, the setting, etc. Some linguistic features (which we will call indicators) show a regular distribution over socioeconomic, ethnic, or age groups, but are used by each individual in more or less the same way in any context. If the social contexts concerned can be ordered in some kind of hierarchy (like socioeconomic or age groups), these indicators can be said to be stratified. More highly developed sogiolinguistic variables (which we will call markers) not only show social distribution, but also stylistic differentiation." See also Labov (t972b:237-240). In connection with the concept "sociolinguistic variable", I will also refer to the concepts "role" and "status"; for a discussion of these concepts see especially Goodenough (1969); for the importance of sociolinguistics for the study of classifier languages, see Barz and Diller (1985) and Diller (1985:56, 64-66).
204
Classificatory Particles in Kilivila
at the beginning of my research. My consultants gave me the "prototypical" noun buda 'group' as the stimulus to elicit these two CP types. In discussing the CP type yuva, I will return to this observation. As Table 3.2.131 also documents, this CP was produced when I expected the CPs gula 'heap, group' and mmwa 'conical bundle' (consultant 45 produced 1 token of the CP buda/boda in each of these cases). The CP type buda/boda/budu normally implies animation of the subjects it refers to; thus, the use of this CP as a response to the stimuli presented to elicit the CPs gula and mmwa is at least somewhat problematic. The CP buda/boda/budu was also produced when I expected bubu 'cut across' and gum 'bit'. These two cases of buda production as a response to the stimuli presented are not acceptable (see also Table 3.2.214). As noted in connection with Table 3.2.131, the CP kwe 'thing' is rather frequently produced by consultants of all five age groups, but especially by consultants of age group IV, instead of the expected CP buda. Table 3.2.43 documents that 3 tokens of the CP type bukwa 'fruit cluster' were produced by 6-year-old Towesei (consultant 10) only. It should be mentioned that Towesei's father Mwasei is highly respected and quite an influential man in Tauwema village. Mwasei, a Malasi, is rather conservative with respect to Trobriand tradition and cultural heritage. As indicated in connection with Table 3.2.132, the CP ke 'wood, long objects, inanimates' obviously supersedes the almost obsolete CP type bukwa. The Tables 3.2.44 and 3.2.133 confirm that the CP type bwa 'tree, wood' plays a secondary role in actual speech (see also Table 3.3.6). It should be noted that the consultants in age group II produce markedly more tokens of this CP type than the consultants in the four other age groups. However, they produce tokens of the CP bwa when I expected utu 'scrap', bubwa 'cut off, kabisi 'compartment/foodhouse', kabulo1 'point', ke 'wood', vili 'untwisted', and si 'bit'. Thus, their definition of the referential frame of this CP is rather broad. Table 3.2.133 shows that we find the production of the CP bwa instead of the expected CPs just mentioned with consultants belonging to other age groups, too. For the sake of completeness, I must mention that the consultants in age group IV produced 3 tokens of the CP bwa instead of the expected CP bubu 'cut across'. It is rather striking that most tokens of the CP bwa were produced when I expected the CPs utu 'scrap' and bubwa 'parts cut off. I elicited only 5 tokens of bwa with the "prototypical" nominal stimulus kai 'tree'. As could have been expected, this stimulus elicited the CP ke 'wood': my consultants gave me the prototypical noun kai as the stimulus to elicit these two CP types. Obviously, ke is the more common of these two CPs. That this is indeed the case becomes even more evident when we compare Table 3.2.44 and the data in appendix A. Twelve of the 21 consultants who produced the CP bwa are members of the Malasi clan (consultants 4, 15, 17, 20, 25, 26, 34, 39, 44, 47, 49, and 60), two consultants are the chiefs children (consultants 28 and 22) and two are his grandchildren (consultants 13 and 19). Moreover, Towesei (consultant 10), whom I already mentioned in connection with the production of the almost obsolete CP bukwa, is also one of the consultants who produced the CP bwa.
The System of Classificatory Particles in Kilivila
205
Thus, 17 of the 21 consultants are of high status in Tauwema. On the basis of these observations, I infer that the CP bwa is a sociolinguistic variable. This finding is confirmed by Tables 3.2.216, 3.2.217, and appendix B. However, the consultants in age group II who produced this CP focused with their responses on the general quality "wooden" of the eliciting noun and not on the eliciting noun phrases, which required more specific classifications. Moreover, the production of the CP bwa instead of the expected vili 'untwisted' seems to indicate that some consultants also mixed up the concept 'rigid long objects', which the CP ke refers to, with the CP bwa, thus extending the frame of reference of bwa. The data I elicited in consultants in age group III indicate that this "overgeneralized" use of a sociolinguistic variable should be restricted again. If this does not happen, the consultants in age group II may change the status of the CP bwa from a sociolinguistic variable to a CP type playing a secondary role in actual speech. Finally it should be noted that this CP type, together with the variant of the CP type ke in its meaning 'wood, wooden things', contradicts Carpenter's (1992:133) remark with respect to "consistent cross-linguistic omission of material". Tables 3.2.45 and 3.2.134 show that the repeater bwalita 'sea' was produced by 4 consultants as a response to the stimulus presented to elicit the CP siwa 'sea portions'. Two of these 4 consultants are of high status in Tauwema: Yebwaku (consultant 38) is a Malasi and Bwetadou (consultant 45) is the weather magician heir. On the basis of this observation, I tentatively infer that although this CP type is becoming obsolete, a few consultants produce it as a sociolinguistic variable. Tables 3.2.46 and 3.2.135 show that although the acquisition of the CP type deli 'group' is somewhat late, this CP nevertheless belongs to the core of the Kilivila CP system. Although the CPs kwe 'thing', kasa 'row', to/te 'male persons, humans' and na1 'female persons' were also produced when I expected the CP deli, the frame of reference of this CP seems to be rather distinct (see Table 3.2.135). However, it must be emphasized that the CP deli generally implies animation of the subject it refers to. Thus, I must note that the CP kwe, the unmarked form for inanimates, here enters the domain of animate subjects. If we look at Table 3.2.166, we realize that the consultants in age groups II and IV are the protagonists of this remarkable development (see also sections 3.2.1.2 and 3.2.1.4) that may eventually cause a loss of distinctness of this CP's frame of reference. Tables 3.2.47 and 3.2.136 document that Bwetadou (consultant 45), the weather magician heir, produced 1 token of the repeater doba 'grass skirt' instead of the expected CP kudu1 'band of fibers'. Table 3.2.124 shows that Bwetadou produces strikingly more CP types than all the other consultants in age group IV. Again, it seems that an almost obsolete CP type is used by this consultant to mark his special status within the village society. Tables 3.2.48 and 3.2.137 confirm that the CP type duli 'cluster, bundle' belongs to the core group of Kilivila CPs. Its frame of reference is rather distinct; nevertheless, we note once more that consultants in age groups I,
206
Classificatory Particles in Kilivila
II, and IV produce the CP kwe 'thing', and that consultants in age group II also produce a few tokens of the CP ke 'inanimates' instead of the expected CP duli. Thus, I note again that the consultants in age groups II and IV foster the frequent use of the CP kwe. As mentioned in section 3.2.1.1, consultant I2/10, Towesei, produced 1 token of the repeater dumia 'swamp' (see Table 3.2.7) in connection with a play accompanying verse transcribed in my overall corpus of Kilivila speech data. Although this CP type seems to be almost obsolete, we should note that it is again Towesei who produces a token of an obviously extremely rare CP (see the previous discussion of the CPs bukwa and bwa). Tables 3.2.49, 3.2.62, 3.2.138, and 3.2.151 document the production of the CP type duya/kaduya/kaduyo 'door, entrance' in its three variants. The variant duya is becoming obsolete. Most consultants produced the CPs kada 'road, track' and ya 'flexible, thin' instead of the expected CP variant duya. However, we also observe the production of a few tokens of the CP types va/vaya/vayo 'door', ke 'wood, long objects, inanimates', and kwe 'thing' instead of the expected CP variant duya. Moreover, some consultants produced the variant kaduyo, and we find 3 tokens of the CP na1 'female persons', an unacceptable response to the stimulus presented to elicit the variant duya (if we exclude the possibility that the informant tried to play a trick on me, using this CP to give the eliciting noun an obscene meaning). The CP variants kaduyo and kaduya are produced more often than the variant duya. However, most consultants produced the CP type kada 'road, track' as a reaction to the stimulus presented to elicit these two variants. Let us set out here on the uncertain terrain of speculation, for a moment, at least. It may be that the variant duya, almost a repeater of the eliciting noun duyava 'door, entrance, window', was superseded by the variants kaduyo and kaduya. The fact that the CP type kada was realized rather frequently instead of the variants kaduyo and kaduya may represent a somewhat similar process. On the basis of this line of speculation, we may assume that kada here is a shortened version of kaduya that is either accidentally homophonous with the CP variant kada of the CP type kada/keda or represents another meaning of the CP type kada/keda. Be that as it may, we also observe the production of the CP types ya 'flexible, thin', ke 'wood, inanimates', kwe 'thing', and va/vaya/vayo 'door' instead of the CP variants kaduya /kaduyo. On the whole, we can record that the CP type duya/kaduya/kaduyo plays a secondary role in actual speech. If we compare Tables 3.2.49 and 3.2.62 with the data in appendix A, the following interesting picture emerges: The tokens of the variant duya were produced by 2 consultants belonging to the highest-ranking Malasi clan (consultants 31 and 55). Three Malasi (consultants 4, 41, and 57), and the chiefs daughter (consultant 50) and granddaughter (consultant 13) are 5 of 7 consultants who produced tokens of the variants kaduya/kaduyo. Thus, we can infer that this CP type, variants of which are produced, with two exceptions, only by consultants of high status, serves the function of a relatively rarely realized sociolinguistic variable. Tables 3.2.50 and 3.2.139 show that the CP gili 'row' is very rarely produced.
The System of Classificatory Particles in Kilivila
207
Most consultants produce the CP ya 'flexible, thin' instead of the expected gili. Moreover, we also observe tokens of the CP types kwe 'thing', kasa 'row', and vili 'untwisted' produced as reactions to the noun that should elicit the CP gill We also notice 1 token of the CP utu 'scrap' that was produced instead of the CP gili; however, this reaction to the stimulus is not acceptable. Again, if we compare Table 3.2.50 with the data in appendix A, we realize that 3 of the 5 consultants producing the CP gili are members of the Malasi clan. The other two consultants are the chief's daughter (consultant 50) and the weather magician heir (consultant 45). Thus, only persons of high status produce a token of this CP type, which obviously also serves the function of a sociolinguistic variable. Tables 3.2.51 and 3.2.140 document that the CP type giwi 'cut' belongs to the core of the Kilivila CP system. Although some tokens of the CPs ke 'wood, inanimates', kwe 'thing', and luva 'wooden dishes' were also produced instead of the expected CP giwi, and although the stimulus presented to elicit this CP elicited also the rather dubious CPs kasa 'row', ta 'basket', and na2 'animals' (see sections 3.2.1.1 and 3.2.1.2), we note that the frame of reference of this CP type is rather distinct. This is quite obvious if we look at the male consultants (see Table 3.2.51). The female consultants not only acquire this CP type later than male consultants, but this CP type also seems to play a somewhat secondary role in their actual language production, especially if we compare their reactions to those of the male consultants in age groups III-V. Tables 3.2.52 and 3.2.141 document that 74-year-old Sedaka (consultant 52, see appendix A) produced 3 tokens of the CP type guba 'bundle of taro'. The production of this CP was not expected. It should be noted that it was, once again, a consultant belonging to the Malasi clan who produced tokens of an extremely rare and almost obsolete CP type. The Tables 3.2.53 and 3.2.142 show that the CP type gudi 'child' belongs definitely to the core of the Kilivila system. Moreover, it is evident that the frame of reference of this CP type is distinct and unequivocal (see also Tables 3.2.216 and 3.2.217). Tables 3.2.54 and 3.2.143 document that the CP type gula 'heap, group' plays a secondary role in actual speech (see also Tables 3.2.216 and 3.2.217). Consultants 17 and 31 also produced a few tokens of the CP gula as a response to the stimulus presented to elicit the CP bubu 'cut across'; however, the use of the gula is not acceptable in this context (see sections 3.2.1.2 and 3.2.1.3). The CP type gula was also produced by consultant 31 as a reaction to the stimulus presented to elicit the CP type po'ula 'grove'. Although this usage is rather idiosyncratic, it is acceptable. Morevoer, Table 3.2.143 shows that tokens of the CP types kwe 'thing', tam 'sprouting', and buda 'group' were also produced as reactions to the stimulus presented to elicit the CP gula. As the data document, the CP kwe has almost superseded the CP gula. Tables 3.2.55 and 3.2.144 show that the CP gum 'bit, small piece' must still be classified as belonging to the core of the Kilivila CP system. However, as Table 3.2.144 shows, a few tokens of the CP types kwe 'thing', ke
208
Classificatory Particles in Kilivila
'inanimates', ya 'thin', and utu 'scrap' were also produced in response to the stimulus presented to elicit the CP type gum. Moreover, a few tokens of the CPs na2 'animals', buda 'group', and kasa 'row' were realized when the CP gum was expected. As explained in sections 3.2.1.1 and 3.2.1.2, the CPs na2 and buda are not acceptable in this context, and kasa is quite dubious. On the basis of these findings, we may conclude that the frame of reference of the CP gum is starting to lose its distinctness and this CP is thus probably losing its status as one of the core types of the Kilivila CP system. Tables 3.2.56, 3.2.120, 3.2.145, and 3.2.209 document the production of the CP type iga/yegila 'name' in its two variants. The variant yegila seems to be almost obsolete; it is produced by only 4 consultants in age group III. The tables confirm that this CP belongs to the core of the Kilivila CP system. Although we find a few tokens of the CP kwe 'thing' produced by 2 consultants as a reaction to the stimulus presented to elicit this CP, and although we note that 1 token of the CP kasa 'row' was produced when the variant iga was expected (a response that is not acceptable; see section 3.2.1.1), we note that the frame of reference of this CP type is distinct. Tables 3.2.58 and 3.2.147 confirm that the CP type kabisi 'compartment of a foodhouse' plays a secondary role in actual speech. As Table 3.2.147 shows, tokens of the CP types he 'wood, inanimates', kwe 'thing', bwa 'tree, wood', utu 'scrap, parts cut off, and pila 'part' were also produced when kabisi was expected. It should be noted that the "general" CP type ke is starting to supersede the more specific CP kabisi. For the sake of completeness, I also should mention that 1 token of the CP nunu 'corner of a garden' was produced by 1 consultant in age group II when the CP kabisi was expected; this usage, however, is not acceptable (see section 3.2.1.2). Tables 3.2.148, 3.2.149, 3.2.59, and 3.2.60 document the production of the CP type kabulo/kabulu. With two different noun phrases I wanted to elicit two (of four) meanings of this CP type, namely: (1) protuberances, points (see Tables 3.2.59 and 3.2.148 ( = kabulo1) and (2) village sectors, areas of authority (see Tables 3.2.60 and 3.2.149 ( = kabulo2)). As these tables show, I could not elicit any token of the CP type kabulo/kabulu in its first meaning with the "prototypical" phrase my consultants mentioned in the test preparatory phase. However, 3 tokens of the CP variant kabulo1 were produced by consultant 49 as a response to the stimulus presented to elicit the CP type nunu 'corner of a garden'. I explained this usage in connection with Table 3.2.148, which shows that tokens of the CPs ke 'wood, long objects, inanimates', utu 'scrap', bwa 'tree, wood', and bubwa 'cut off were realized instead of tokens of the variant to be elicited. The production of all these CPs is acceptable, and this is obviously due to the consultants understanding of the eliciting phrase, which does not seem to be as "prototypical" as the consultants in the test preparatory phase claimed. As to the second variant of kabulo/kabulu, Tables 3.2.149 and 3.2.60 show that kabulo2 plays a secondary role in actual speech (see also Tables 3.2.216 and 3.2.217). However, if we compare Table 3.2.60 and appendix A, we note that 6 of the 11 consultants producing this CP type variant are members of
The System of Classificatory Particles in Kilivila
209
the highest-ranking Malasi clan (consultants 24, 26, 31, 34, 49, and 60), that consultants 28 and 53 are the chief's daughter and wife, that consultant 45 is the weather magician heir, and that consultant 46 is one of the most respected navigators in Tauwema. Thus, 10 of 11 consultants producing this CP are of high status in the Tauwema village community. Therefore, I infer that the CP type kabulo/kabulu, at least in its second meaning ( = kabulo2), is a sociolinguistic variable. This is also confirmed by Tables 3.2.216 and 3.2.217 and appendix B. Table 3.2.149 also shows that the CP type ke 'wood, long objects, inanimates' and, though to a lesser degree, the CP type kwe 'thing, inanimates' have superseded the CP variant kabulo2. We also find 1 token of the CP utu 'scrap' produced when kabulo2 was expected. Moreover, 3 tokens of the CP kada 'road' were also produced instead of the expected CP variant; however, kada in this context is not acceptable (see section 3.2.1.3). We conclude that the CP type kabulo/kabulu plays a secondary role in actual speech and is superseded by the general CP types ke and kwe. If this CP is produced, however, at least in its second meaning (kabulo2), it serves the function of a sociolinguistic variable. Tables 3.2.61 and 3.2.150 document the production of the CP kada/keda 'road, track', which obviously belongs to the core of the Kilivila CP system (see also Tables 3.2.216 and 3.2.217). Its frame of reference encompasses not only the expected frame but also, to a rather large degree, the frame of reference of the CP type duya/kaduya/kaduyo 'door, entrance' (see the previous discussion of this CP type). A few tokens of the CP type kada/keda were also produced when the CP types lilo 'walk', vilo 'place', and kabulu/kabulo 'village sector' (= kabulo2) were expected. However, the use of the CP type in these contexts is either dubious or not acceptable (see sections 3.2.1.2 and 3.2.1.3). Tables 3.2.63 and 3.2.152 document the production of the CP type kai 'stone blade', one of the core types of the Kilivila CP system (see also Tables 3.2.216 and 3.2.217). The tables show that the frame of reference of this CP type is distinct. We only find 2 tokens each of the synonymous repeater beku 'stone blade' and of the general CP type kwe 'thing' produced instead of the expected CP kai. The production of 1 token of the CP variant na1 'female persons' and of 2 tokens of the CP type sisi 'bough' produced instead of the expected CP kai is not acceptable in this context (see section 3.2.1.2). Tables 3.2.64 and 3.2.153 document the production of the CP type ka'i 'tooth', which was not expected. Nevertheless, it belongs to the core of the Kilivila CP system. As Table 3.2.153 shows, the CP ka'i was produced when the synonymous CP variant kudu2 'tooth' was expected. Obviously, the CP ka'i has superseded the CP variant kudu2 (see also the later discussion of the CP type kudu). We note 3 tokens of the CP ka'i produced as a response to the stimulus presented to elicit the CP oyla 'string'. However, as stated in section 3.2.1.2, the use of the CP ka'i in this context is not acceptable. Thus, we conclude that the CP ka'i has superseded the synonymous CP variant kudu2 and that the frame of reference of this CP is distinct.
210
Classificatory Particles in Kilivila
Tables 3.2.65 and 3.2.145 and appendix A show that only Weyei (consultant 55), the 62-year-old weather magician and the chief's brother, produced 3 tokens of the CP kala 'day'. The general CP type kwe 'thing' has obviously superseded this almost obsolete CP type. For the sake of completeness, I mention that the consultants produced 1 token of the repeater tuta 'time' and 5 tokens of the repeater yam 'day' when the CP kala was expected. Again, I have to note that a consultant of status (Weyei is also a member of the highest-ranking Malasi clan) produced an extremely rare and almost obsolete classifier. In my recorded and transcribed texts, there is 1 token of the CP type kali 'paddle strike" realized by 60-year-old Mokopei (consultant V14) telling a Trobriand mythical story (see Tables 3.2.216 and 3.2.217 and appendix B). Discussing this CP with my consultants, I found that this CP type is almost obsolete and only preserved in important texts of oral tradition. Tables 3.2.66 and 3.2.155 document the production of the CP kapwa/kapo 'parcel, bundles (wrapped up)'. This CP type plays a secondary role in actual speech. It is superseded by the CP luba 'bundle'. Moreover, 30 tokens of the general CP kwe 'thing' were produced by consultants of all five age groups when the CP kapwa/kapo was expected. If we compare Table 3.2.66 with the data in appendix A, we note the following: 8 of the 18 consultants who produced tokens of the CP kapwa/kapo are members of the highest-ranking Malasi clan (consultants 11, 15, 20, 25, 30, 47, 49, and 55). Five of the other 10 consultants are either the chief's children (consultants 22, 28, and 50) or his grandchildren (consultants 13 and 19). Towesei (consultant 10), whom we already mentioned in connection with the CP types bukwa and bwa, is among the five remaining consultants. Thus, we record that 13, or, if we include Towesei, 14 of the 18 consultants that produced the CP kapwa/kapo are of high status within the Tauwema community. We may infer from this observation that the CP kapwa/kapo also functions as a sociolinguistic variable. Tables 3.2.67 and 3.2.156 document the production of the CP kasa 'row, line' which belongs to the core of the Kilivila CP system (see also Tables 3.2.216 and 3.2.217). We also note that a few tokens of this CP were produced when the CP types meila 'part of a song', gili 'row', nina 'part of a song', and deli 'group on the move' were expected. With one consultant (consultant 23), we find 3 cases of dubious realization of the CP kasa instead of the expected CPs giwi 'cut', gum 'bit', and yuva 'shoal' (see section 3.2.1.2). Moreover, consultant 6 produced 1 token of the CP kasa instead of the expected CP iga 'name'; the use of kasa is not acceptable in this context (see section 3.2.2.1). As Table 3.2.156 indicates, we only find a few tokens of the general CP types ke 'inanimates' and kwe 'thing', which were produced when kasa was expected. Thus, we note that the frame of reference of this CP is distinct. In my recorded and transcribed texts, I find 8 tokens of the CP type kauya 'fish trap, creel' produced by Kalavatu (consultant IV3) and Tokunupei (consultant VI5) in their description of their special skills as the basket makers in Tauwema. Thus, we note that this repeater belongs to the technical language variety of Trobriand wicker workers.
The System of Classificatory Particles in Kilivila
211
Tables 3.2.69 and 3.2.158 and appendix A document that 75-year-old Vaka'ila, a member of the highest-ranking Malasi clan, is the only consultant who produces 3 tokens of the CP type kavi 'tool'. Table 3.2.158 also documents that the noun presented to elicit this repeater only elicited the production of a few tokens of the general CPs ke 'inanimates' and kwe 'thing' and the dubious and unacceptable production of 1 token each of the CPs pila 'part' and utu 'scrap'. What is striking here is that only consultants in age group II produced these responses to the eliciting noun. We may interpret this as an indication of the children's high motivation to acquire the Kilivila CP system. In any case, we can assume that the CP type kavi is becoming obsolete. Tables 3.2.70 and 3.2.159 document the production of the CP ke 'wood, long objects, inanimates, fire'. The exceptional status of this CP type is obvious (see also Tables 3.2.216 and 3.2.217). It is not only one of the first CP types acquired, but its frame of reference with respect to the meaning 'rigid, long objects' and its function as one of the two "unmarked forms for inanimates" is very broad. However, its frame of reference with respect to the meanings 'wooden things' and 'fire' is distinct. Only 2 tokens of the CP type bwa 'tree, wood' were produced instead of the expected CP ke. Moreover, Table 3.2.159 documents that the CP ke has superseded the CP bwa. However, this is also true for the CP types lila 'bough, branch, leaf and luva 'wooden dishes'. The CP ke has almost superseded the CP types bubwa '(parts) cut off, bukwa 'fruit cluster', kabulo/kabulu 'point, sector', kila 'hands of bananas', mmwa 'conical bundle', and sa 'nut bunch'. The fact that so many tokens of the CP type ke were produced instead of the expected CP uva 'span', however, is obviously a result of the consultants' focussing on the 'yams' mentioned in the eliciting noun phrase instead of referring to the concept expressed by the eliciting noun phrase as a whole. We also find tokens of the CP type ke produced by consultants of all five age groups instead of the expected CPs utu 'scrap', kabisi 'compartment of a foodhouse', giwi 'cut', and ya 'flexible, thin'. However, as the data document, the CP ke plays a secondary role as an acceptable substitute for these CP types. This is also true for the production of the CP ke instead of the expected CPs liku 'compartment of a foodhouse', vili 'untwisted', si 'bit', bubu 'cut across', oyla 'string', sisi 'bough', meila 'part of a song', gum 'bit' duya/kaduya 'door, entrance', kubila 'land plot', kwoila 'clay pot', pila 'part', yulai 'bundle of four things', kweya 'limb', kudu2 'tooth', kavi 'tool', duli 'cluster', yuva 'shoal', sipu 'sheaf, lilo 'walk', and kasa 'row'. We also find only 1 token of the CP ke in its meaning 'fire' (= ke4) produced instead of the expected, synonymous repeater kova 'fire'. As mentioned in connection with Table 3.2.159, it should be emphasized that consultant 12 produced 1 token of the CP ke instead of the expected CP yuva 'shoal', which normally implies animation of the subjects it refers to. However, whether this can be regarded as a first indication that the CP type ke is starting to invade the semantic domains characterized by the feature 'animation' must remain an open question here (see section 3.3.4). Tables 3.2.70 and 3.2.159 also show that the consultants in age group II produced by far the most tokens of this CP type. If we list the number of cases where ke was produced instead of the expected CP type, we find the
212
Classificatory Particles in Kilivila
following: The consultants in age group II produced tokens of the CP ke for 29 different expected CP types, 3 more than we find with the consultants in age group I. However, the older consultants in age groups IV and V also produced tokens of the CP ke for 24 and 23 different expected CP types, respectively. The consultants in age group III produced tokens of the CP ke for 19 different expected CP types. These data suggest the following interpretation: Consultants in age group II produced the CP ke in its broadest frame of possible reference. On the other hand, the consultants in age group III seem to be more selective in, and aware of, their choice of more specific CPs within the Kilivila CP system (see also Table 3.2.124). It goes without saying that we must regard this interpretation with all necessary caution. What we can conclude, however, is that, on the one hand, the frame of reference of the CP type ke in its meanings 'wooden things' and 'fire' is distinct, and, on the other hand, the frame of reference of this CP in its meaning 'rigid long objects', and especially in its function as one of the unmarked forms for inanimates, is so broad that we can assign it the status of a so-called general classifier, especially the last mentioned CP variant (cf. Adams 1989:182). Tables 3.2.71 and 3.2.160 document the production of the CP type kila 'clusters/hands of bananas', We note 7 tokens of this CP type realized by 4 women only. Table 3.2.71 and the data in appendix A document that these women are all members of the highest-ranking Malasi clan. As mentioned previously in connection with the CP type ke, Table 3.2.160 shows that ke has superseded the CP type kila. We also find a few tokens of the CP type kwe 'thing' produced instead of the expected CP kila. Thus, we can conclude that the CP kila is becoming obsolete; however, it seems to function as a sociolinguistic variable for high-ranking Trobriand women who still use the distinct frame of reference of this CP as a special mass term. Tables 3.2.72 and 3.2.161 document the production of the CP type kova 'fire', which belongs to the core of the Kilivila CP system. Although we notice that consultant 5 produced 1 token of the CP kova instead of the expected CP yuma 'hand' (a use of kova that is not acceptable; see section 3.2.1.1), and although 1 token of the CP kwe 'thing' was produced instead of the expected CP kova, we can state that the frame of reference of this repeater is distinct (see also Tables 3.2.216 and 3.2.217). It should also be noted that Table 3.2.161 documents 1 token of the CP type ke in its meaning 'fire', produced instead of the synonymous repeater kova (see the previous discussion of the CP ke). Tables 3.2.73 and 3.2.162 document the production of the CP type kubila/kwabila 'large land plot', which also belongs to the core of the Kilivila CP system. It seems that this CP is somewhat more important in the language of men than in that of women, thus reflecting the division of labor in Trobriand society (see, e.g., Malinowski 1929, 1935; Hutchins 1980). Table 3.2.162 also shows that only 2 tokens of the CP type utu 'scrap' were produced instead of the expected CP kubila/kwabila; this use of the CP utu, however, is not
The System of Classificatory Particles in Kilivila
213
acceptable. We also note that the consultants produced 1 token each of the CP types ke 'inanimates' and 'pila 'part' instead of the expected CP kubila/ kwabila. What is striking, however, is that this CP type with its distinct frame of reference seems to also be in danger of being superseded by the CP type kwe 'thing': there were 53 tokens of the CP kwe produced by consultants of all five age groups instead of the expected specific CP. The fact that in my recorded and transcribed texts only 1 token of the CP kubila/kwabila was produced may be a result of the text categories recorded and transcribed (see Tables 3.2.216 and 3.2.217 and appendix B). Tables 3.2.74, 3.2.75, 3.2.163 and 3.2.164 document the production of the CP type kudu in its meanings 'band of fibers' (kudu1) and 'tooth' (kudu2). As Tables 3.2.74 and 3.2.163 show, the variant kudu1 'band of fibers' is becoming obsolete (see also Tables 3.2.216 and 3.2.217). Moreover, consultant 58 produced tokens of this CP variant when the CP types bubu 'cut across' and yuva 'shoal' were expected; all my other consultants agreed that the use of the variant kudu1 is not acceptable in either context (see section 3.2.1.5). Table 3.2.163 also shows that this CP variant has been superseded by the CP kwe 'thing'. We also notice the production of the repeater doba 'grass skirt' as a reaction to the stimulus presented to elicit the variant kudul (see the previous discussion of the repeater doba). Tables 3.2.75 and 3.2.164 document that the CP variant kudu2 'tooth' plays only a secondary role in actual speech. However, if we compare Table 3.2.75 and the data in appendix A, we realize that 9 of the 12 consultants producing this CP variant belong to the highest-ranking Malasi clan. Thus, we may infer that the variant kudu2 functions as a sociolinguistic variable. In the previous discussion of the CP ka'i I mentioned that this CP type has almost superseded the variant kudu2. For the sake of completeness, I should also mention that the consultants produced 2 tokens each of the CP types ke 'long objects, inanimates' and kwe 'thing' instead of the expected CP variant kudu2. Tables 3.2.76 and 3.2.165 document the production of the CP type kumla 'earth oven', one of the core types of the Kilivila CP system. It is quite obvious that the frame of reference of this CP is distinct. Only 1 token of the CP kwe 'thing' was produced instead of the expected CP kumla. Tables 3.2.77 and 3.2.166 document the production of the CP type kwe 'thing', unmarked form for inanimates, clams, and shells'. Again, the exceptional status of this CP is obvious (see also Tables 3.2.216 and 3.2.217). The broad range of meaning that is covered by this CP type is clearly documented in Table 3.2.166. The CP kwe has superseded the CPs bogi 'night', kala 'day', kudu1 'band of fibers', kwoya/koya 'mountain', lilo 'walk', siva 'time', siwa 'sea portions', vilo place', yam 'day', and yulai 'bundle of four things'. It has almost superseded the CP types gula 'heap', nunu 'corner of a garden', po'ula 'grove', and 0 'basketful of yams'. The fact that all consultants produced tokens of the CP kwe instead of the actually expected CPs buda 'group', kapwa 'parcel', kubila 'land plot', kwoila 'clay pot', nina 'part of a
214
Classificatory Particles in Kilivila
song', sipu 'sheaf, yuva 'shoal', and duya/kaduyo 'door, entrance' may indicate that these CPs are also in danger of being superseded by this classifier. The same seems to hold for the CPs bubu 'cut across', gili 'row', kabulo2 'sector', deli 'group', kabisi 'compartment of a foodhouse', gum 'bit', liku 'compartment of a foodhouse', notu 'kneaded, dot', 'pwanina/pona 'hole', sa 'nut bunch', and yuma 'hand'. The fact that we find the CPs buda, deli, and yuva among this group is worthy of note. These CPs normally imply animation of the subjects they refer to. On the basis of the data presented, we can infer that the CP kwe has started to invade the semantic domains characterized by the feature 'animate' (for a similar situation in Hokkien, see Chin 1989:37-39). The fact that we find tokens of the CP kwe produced instead of the expected CPs giwi 'cut', vili 'untwisted', duli 'cluster', kila 'hands of bananas', meila 'part of a song', oyla 'string', mmwa 'conical bundle', iga/yegila 'name', kasa 'row', kudu2 'tooth', bukwa 'fruit cluster', uva 'span', kova 'fire', kumla 'earth oven', pila 'part', bubwa 'cut off, ta/peta 'basket', kai 'stone blade', kavi 'tool', kweya 'limb', luva 'wooden dishes', and si 'bit' documents the enormous possibilities of an acceptable usage of this CP type. We can attribute to the CP kwe the status of a so-called general classifier. A closer look at Table 3.2.77 reveals that the consultants in age groups II and IV especially seem to foster the development of the CP kwe from the 'unmarked form for inanimates' to a general classifier. This observation is at least in part confirmed if we look at the data presented in Table 3.2.166, which document the production of the CP kwe instead of the expected CPs buda, yuva, and deli. Here, too, we observe that it is the consultants in age groups II and IV that foster the invasion of the CP kwe into the semantic domains characterized by the feature 'animate'. In connection with this observation, we should emphasize, too, that kwe is also the adequate CP to refer to 'shells and clams' (see Tables 3.2.216 and 3.2.217). In my recorded and transcribed texts, Toybokwatauya (consultant IV1/43) and Kilagola (consultant V11) produce 2 tokens each of the CP type kwe in this meaning. Mokopei (consultant V14) produced 15 tokens of the CP kwe in its meaning 'shells and clams'. The corpus on Trobriand myths offers an interesting, though highly speculative, possibility to explain this meaning of the CP kwe, which implies also an exception in the frame of reference of the CP na in its meaning 'animals (but not shells and clams)'. One of the most important mythical folktales is the story of Imdeduya and Yolina (see, e.g., J. W. Leach 1981; Kasaipwalova 1980). I recorded and transcribed three different versions of this folktale, the longest and most elaborate version of which was told to me by Mokopei from Kaduwaga village. I assume that Imdeduya and Yolina are the personifications of the moon and sun. Yolina had to paddle a long distance in his canoe before he came to the place where Imdeduya lived. Mokopei mentioned that the canoe of Yolina was a kweduya; this is the archaic name for a shell that is now called duya (i.e., Melo aethiopicus [Linnaeus] of the family Volutidae). It
The System of Classificatory Particles in Kilivila
215
may be that the CP kwe in its meaning 'clams and shells' is a mythical relic, a honorific and revertential reminiscence of Yolina's excellent ocean-going canoe. I want to emphasize again that; this attempt to explain this meaning of the CP kwe, is of course, highly speculative; however, on the basis of my Trobriand experience, it is quite plausible.5 Tables 3.2.78 and 3.2.167 document the production of the CP kweya/kwaya/ keya '(severed) limb', one of the core types of the Kilivila CP system. The frame of reference of this CP type is distinct (see also Tables 3.2.216 and 3.2.217). As the more general CP type, it seems to supersede the CP type yuma 'hand', as Table 3.2.167 shows. We find only a few tokens of the general CPs ke 'inanimates' and kwe 'thing' produced instead of the expected CP kweya. Table 3.2.167 also shows 4 tokens of the CP variant na1 'female persons' that were produced instead of the expected CP kweya', however, the use of this CP variant in the given context was not acceptable (see sections 3.2.1.1 and 3.2.1.2). Tables 3.2.79 and 3.2.168 document the production of the CP kwoila/kwela 'clay pot', which belongs to the core of the Kilivila CP system. The frame of reference of this CP type is distinct (see also Tables 3.2.216 and 3.2.217). However, as Table 3.2.168 shows, the CP kwe is starting to supersede this CP. We also note the production of a few tokens of the CP ke 'inanimates' instead of the expected CP kwoila/kwela. Consultant 58, Kasilasila, also produced one token of *kava in this context; however, according to all my other consultants, such a CP does not exist in Kilivila (see Tables 3.2.68 and 3.2.157); I can only assume that *kava was an error of speech production. Tables 3.2.80 and 3.2.169 document the production of the CP kwoya/koya 'mountain, hill', which plays a secondary role in actual speech. The frame of reference of this CP is distinct; however, the CP kwe 'thing' has superseded it. If we compare Table 3.2.80 and the data in appendix A, we find the following: 5 of the 13 consultants who produced the CP kwoya/koya are members of the highest-ranking Malasi clan (consultants 1, 15, 31, 34, and 41); 4 of the other consultants belong to the chief's family (consultants 21, 32, 28, 53), and consultant 45 is the weather magician heir. Thus, 10 of the 13 consultants producing tokens of this CP type are of high status in Tauwema. Hence, we may infer that the CP kwoya/koya functions as a sociolinguistic variable. Tables 3.2.81 and 3.2.170 document the production of the CP liku 'compartments of a foodhouse/of a canoe', which plays a secondary role in actual speech. Its frame of reference is distinct (see also Tables 3.2.216 and 3.2.217); however, it is in danger of being superseded by the CP kwe 'thing' and also by the CP ke 'wood, inanimates'. If we compare Table 3.2.81 and the data in appendix A, we find the following: 27 consultants produced tokens of the CP liku; 14 of these consultants are members of the highest-ranking 5 In connection with this speculation. Bousfield (1979:217) stated, "Any theoretical account must first deal with the complicated, untidy logics actually found". See also Chin (1989:33-34); Dixon (1972:302-304).
216
Classificatory Particles in Kilivila
Malasi clan (consultants 15, 17, 24, 26, 30, 34, 38, 39, 49, 51, 55, 56, 59, and 60); 6 consultants belong to the chief's family (consultants 19, 21, 32, 28, 43, and 48), and 1 consultant is the weather magician heir (consultant 45). Thus, 21 of the 27 consultants who produced tokens of the CP liku are of high status, and we may assume that the CP liku functions as a sociolinguistic variable (see also Tables 3.2.216 and 3.2.217 and appendix B). As mentioned in section 3.2.2.2, the CP type lila 'bough, branch, leaf was not produced by my consultants during the elicitation test; nor was it produced in the recorded and transcribed speech data (see Tables 3.2.216 and 3.2.217). Instead of the expected CP lila, consultants of all five age groups produced 125 tokens of the CP ke 'wood, long objects, inanimates' in the elicitation test. Thus, although the consultants claimed during the informal pretests in the test preparatory phase that this CP type is well known and used in their speech community, it seems that the CP lila is either obsolete or extremely rare and that the CP ke has superseded it. Tables 3.2.82 and 3.2.171 document the production of the CP lilo/lola 'walk, journey'. This extremely rare and almost obsolete CP type has been superseded by the CP kwe 'thing'. We also note some tokens of the CP ya 'flexible, thin' produced instead of the expected lilo/lola. My consultants also produced 1 token each of the CPs ke 'inanimates' and kada 'road' instead of the expected lilo/lola; the use of the CP kada in this context, however, is somewhat dubious (see section 3.2.1.2). If we compare Table 3.2.82 and the data in appendix A, we see that 2 of the 4 consultants producing the CP lilo/lola are members of the Malasi clan and that the weather magician heir also produced 1 token of this CP type. However, the database is too small to decide whether we can regard the CP lilo/lola as a sociolinguistic variable. As documented in Tables 3.2.216 and 3.2.217 and appendix B, Kalavatu (consultant IV3) produced tokens of the CP type lipu 'compartment of a creel'. As mentioned in connection with the previous discussion of the CP kauya 'creel', Kalavatu is one of the basket makers of Tauwema village. Thus, we can infer that this extremely rare CP type belongs to the technical language variety used by Trobriand wicker workers. Tables 3.2.83 and 3.2.172 document the production of the CP luba 'bundle'. This repeater, which was not expected, plays a secondary role in actual speech. However, it has superseded the CP kapwa 'parcel, bundle'. Tables 3.2.84 and 3.2.173 document the production of the CP luva 'wooden dishes, tied bundle', which plays a secondary role in actual speech production. As Table 3.2.173 indicates, only 4 tokens of the variant luva1 'wooden dishes' were produced by consultants belonging to age groups I and V. We note 3 tokens each of the CP variant luva2 'tied bundle' produced by consultants belonging to age group IV, instead of the expected CPs mmwa 'conical bundle' and sa 'nut bunch', and 2 tokens of this CP variant produced by a speaker of the same age group instead of the expected CP bukwa 'fruit cluster'. The production of the variant luva2 was not elicited. As noted in Table 3.2.173, the production of the CP luva instead of the expected CPs utu 'scrap' and giwi 'cut' is not acceptable. The CP ke 'wood, inanimates' has superseded
The System of Classificatory Particles in Kilivila
217
the CP luva. We also find 2 tokens of the CP kwe 'thing' produced instead of the expected CP luva. It seems that the CP luva is becoming obsolete. Tables 3.2.85 and 3.2.174 document the production of the CP megwa 'magic, magical formula'. The production of this extremely rare repeater was not expected. Only 7 consultants produced tokens of this CP instead of the expected CP nina 'part of a song'. Although a comparison of Table 3.2.85 and the data in appendix A shows that 1 of the 7 consultants is a member of the highest-ranking Malasi clan (consultant 38), that 2 other consultants belong to the chiefs family (consultants 24 and 38), and that 1 consultant is the weather magician heir (consultant 45), it is not possible to determine whether this CP type functions as a sociolinguistic variable or belongs to the technical language variety of Trobriand magicians. Tables 3.2.86 and 3.2.175 document that only Weyei (consultant 55) produced 3 tokens of the extremely rare and almost obsolete CP meila 'part of a song/a magic formula'. The CP pila 'part' has obviously superseded the CP meila. We also note the production of some tokens of the CPs vosi 'song', ke 'inanimates', kasa 'row', kwe 'thing', and utu 'scrap, parts' as responses to the stimulus presented to elicit the CP meila. Tables 3.2.87 and 3.2.176 document the production of the CP mmwa 'conical bundle (most often of taro)', which plays a secondary role in actual speech. The CP ke 'inanimates' has almost superseded it. We also note the production of a few tokens of the CPs kwe 'thing', guba 'bundle of taro', buda 'group' and luva2 'tied bundle', and utu 'scrap, part' instead of the expected CP mmwa. The production of the CP luva2 in this context is a bit dubious, and the production of the CP utu in this context is not acceptable (see sections 3.2.1.3 and 3.2.1.4). If we compare Table 3.2.87 and the data in appendix A, we find the following: 7 of the 10 consultants who produced tokens of the CP mmwa are members of the highest-ranking Malasi clan (consultants 31, 34, 41, 49, 51, 55 and 57) and 1 consultant is the chiefs grandson (consultant 19). Thus, 8 of these 10 consultants are of high status. Only 3 of these 8 consultants are women. We may tentatively conclude that the CP mmwa functions as a sociolinguistic variable and that this function seems to be important especially for Trobriand men. Tables 3.2.88, 3.2.89, 3.2.177, 3.2.178, 3.2.216, and 3.2.217 document the production of the CP type na in 4 of its 6 meanings. With its meanings 'persons of female gender' (na1) and 'animals, but not shells and clams' (na2), this CP type belongs to the core of the Kilivila CP system. As Tables 3.2.216 and 3.2.217 indicate, the CP na is not often produced in its meanings 'stars, planets, moon' (na3) and 'carvings in human likeness' (na4). Tables 3.2.88 and 3.2.177 show that the variant na1 'female persons' is acquired very early; its frame of reference is distinct. With consultants of the first two age groups we also note the production of a few tokens of this CP variant instead of the CPs kweya 'limb', bubu 'cut across', uva 'span', kai 'stone blade', and deli 'group'. With the exception of deli, the production of the variant na1 as a substitute for these CPs mentioned is not acceptable in the respective elicitation contexts (see sections 3.2.1.1 and 3.2.1.2). One
218
Classificatory Particles in Kilivila
consultant of age group III, Namnabai (consultant 25), produced 3 tokens of the variant na1 as a response to the stimulus presented to elicit the CP duya 'door, entrance'. If we exclude a possibly metaphorical, though rather obscene, extension of the meaning 'person of female gender' of the variant na1, we have to record that the use of na1 instead of the expected CP duya is not acceptable (see also section 3.2.1.3). Tables 3.2.89 and 3.2.178 show that the frame of reference of the CP variant na2 'animals, but not shells and clams' (see the previous discussion of the CP kwe) is also distinct. Table 3.2.178 shows that the variant na2 was also produced when the CPs oyla 'string, fish on strings' and wela 'batch of fish' were expected; these CPs refer specifically to caught fish. A few tokens of the CP variant na2 were also produced instead of the expected CPs gum 'bit' and giwi 'cut'; these were not acceptable. However, it may be that the consultant mistook the stimulus giwi 'cut' for the CP type givi, which refers to 'a serving of fish' (see section 3.3.1, Table 3.3.2, L15); such a misunderstanding would explain the production of the CP variant na2, which would then be acceptable. For the sake of completeness, I should mention that my consultants also gave the meanings 'corpses' (na5) and 'spirits, dwarfs' (na6) for the CP type na; however, these variants are not documented in my corpus of recorded and transcribed speech data. Moreover, I would like to mention here that in the Mayan language Mixtec there is also a classifier that refers not only to animals, but also to the rainbow, the moon, and the stars (see de Leon 1988:146). Tables 3.2.90 and 3.2.179 and appendix A document that 62-year-old Weyei (consultant 55) produced only 1 token of the CP type nigwa 'hole' instead of the expected CP type pwanina/pona 'punctured, hole'. The production of this almost obsolete CP was not expected. Tables 3.2.91 and 3.2.180 and appendix A document that the same consultant produced only 3 tokens of the CP type nina 'part of a song', which is almost obsolete. It has been superseded by the CP types pila 'part' and kwe 'thing'. As Table 3.2.179 also shows, we find a few tokens of the CP types megwa 'magic', utu 'scrap, part', and kasa 'row' instead of the expected CP nina. We also note 1 token of a formative *iki produced instead of the expected CP nina. All my other consultants told me that such a CP type does not exist in Kilivila (see Tables 3.2.146 and 3.2.57). However, as noted in connection with Table 3.2.146, it may be that the consultant wanted to produce the CP type igi 'wind' (see Table 3.3.2, L19), a rather idiosyncratic, although understandable response to the stimulus presented to elicit the CP nina. Tables 3.2.92 and 3.2.181 document the production of the CP nutu/notu 'kneaded, dot, drop', which belongs to the core of the Kilivila CP system. The frame of reference of this CP type is distinct. Table 3.2.181 also shows that consultants in four of the five age groups produced 33 tokens of the general CP kwe 'thing' instead of the expected CP nutu/notu. Thus, we may infer that the CP nutu/notu is also in danger of being superseded by the CP kwe.
The System of Classificatory Particles in Kilivila
219
Tables 3.2.93 and 3.2.182 document the production of the CP nunu 'corner of a garden', which plays a secondary role in actual speech. As mentioned in connection with Table 3.2.182, the CP kwe 'thing' has almost superseded nunu. We also note a few tokens of the CPs utu 'scrap, part' and kabulo2 'sector' produced instead of the expected CP nunu. The production of 1 token of the CP kabisi 'compartment of a foodhouse' as a reaction to the stimulus presented to elicit the CP nunu is not acceptable (see section 3.2.1.2). If we compare Table 3.2.93 and the data in appendix A, we find the following: 6 of the 9 consultants who produced tokens of the CP type nunu are members of the highest-ranking Malasi clan (consultants 15, 26, 38, 39, 55 and 56) and consultant 28 is the chief's daughter. Thus, 7 of the 9 consultants who produced this CP type are of high status in Tauwema. We may tentatively conclude therefore that the CP type nunu may also function as a sociolinguistic variable. In addition, it should be noted that the frame of reference of this CP type is distinct. As mentioned at the end of section 3.2.2.2, the CP type oyla 'string, fish on strings' was neither produced by my consultants during the elicitation test nor found in the recorded and transcribed texts (see Tables 3.2.216 and 3.3.217). Instead of the expected CP type oyla, the consultants produced tokens of the CPs na2 'animals', ke 'long objects, inanimates', wela 'batch of fish', and kwe 'thing'. We also find a few tokens of the CPs ka'i 'tooth' and pila 'part' produced instead of the expected CP oyla. However, the two last mentioned responses to the stimuli presented to elicit oyla are not acceptable or are rather dubious. Thus, although my consultants claimed during the informal pretests that this CP type is well known and used in their speech community, it seems that oyla is either obsolete or extremely rare and that the CPs na2 'animals', ke 'long objects, inanimates', and wela 'batch of fish' have superseded it. Tables 3.2.94, 3.2.104, 3.2.183, and 3.2.193 document the production of the two variants of the CP type peta/ta 'basket'. The variant ta obviously belongs to the core of the Kilivila CP system, whereas the repeater peta is extremely rare or almost obsolete. If we compare Tables 3.2.94 and 3.2.104, we note that the 6 consultants who produced the variant peta produced it together with the variant ta. It may be that the repeater functions as a sociolinguistic variable: 3 of the 6 consultants producing the variant peta are members of the highest-ranking Malasi clan (consultants 27, 57, and 60), consultant 50 is the chief's daughter, and consultant 45 is the weather magician heir (see appendix A); however, the database is rather small for such a hypothesis. We only note one dubious use of the CP ta: 1 token of the variant was produced instead of the expected CP giwi 'cut' (see section 3.2.1.1). Moreover, we only observe the production of 1 token of the CP type kwe 'thing' instead of the expected variant ta. Thus, we conclude that the frame of reference of the CP type peta/ta is distinct (see also Tables 3.2.216 and 3.2.217). Moreover, Table 3.2.104 confirms the observation that this CP type is acquired rather early. Tables 3.2.95 and 3.2.184 document the production of the CP pila 'part', which belongs to the core of the Kilivila CP system. Table 3.2.184 records
220
Classificatory Particles in Kilivila
the rather broad frame of reference of this CP type (see also Tables 3.2.216 and 3.2.217). The CP pila has started to supersede or has almost superseded the CPs nina 'part of a song' and meila 'part of a song'. Moreover, we also find a few tokens of the CP pila that were produced instead of the CPs ya 'flexible, thin', siva 'time', kabisi 'compartment of a foodhouse', and kubila 'land plot'. The few tokens of the CP pila that were produced instead of the expected CPs oyla 'string' and kavi 'tool', however, document dubious uses of this CP type (see the previous discussion of the CP oyla). We also note that consultants of all five age groups produced 60 tokens of the CP ya 'flexible, thin' instead of the expected CP pila; the CP ya seems to cover a rather important part of the obviously broad frame of reference defined for the CP pila. A few tokens of the CPs ke 'wood, inanimates', kwe 'thing', utu 'scrap', and vili 'untwisted' were also produced instead of the expected CP pila. Tables 3.2.96 and 3.2.185 record the production of the CP po'ula 'plantation, grove', which plays a secondary role in actual speech. Although Table 3.2.185 documents that a few tokens of this CP were also produced instead of the expected CP types sa 'nut bunch' and bukwa 'fruit cluster' (a completely acceptable response to the stimuli presented to elicit these two types), it should be noted that the frame of reference of the CP po'ula is distinct. However, as mentioned previously, the CP po'ula has been almost superseded by the CP kwe. The fact that we find 3 tokens of the CP gula 'heap, group' produced instead of the CP po'ula is not significant; it is mentioned here only for the sake of completeness. Tables 3.2.97 and 3.2.186 document the production of the two variants of the CP type pwanina/pona 'punctured, hole', which belongs to the core of the Kilivila CP system. As noted in Table 3.2.97, the ratio of production for the two variants pwanina and pona is 94:45. Thus, with this CP type, the repeater variant is still preferred by my consultants. Table 3.2.186 documents that the frame of reference of this CP type is distinct (see also Tables 3.2.216 and 3.2.217). We only note 23 tokens of the CP kwe 'thing' and 1 token of the CP nigwa 'hole' produced instead of the expected CP pwanina/pona. Tables 3.2.98 and 3.2.187 record the production of the CP sa 'nut bunch', which is obviously extremely rare or almost obsolete. Although 3 tokens of the CP sa were produced instead of the expected CP bukwa 'fruit cluster', the frame of reference of the CP sa is distinct. As also documented in Table 3.2.187, the CP type ke 'wood, inanimates' has almost superseded sa. Moreover, we also find a number of tokens of the CP kwe 'thing' and a few tokens of the CPs po'ula 'grove' and luva2 'tied bundle' instead of the CP sa. The production of the CP luva2 instead of the CP sa is somewhat dubious, however (see section 3.2.1.4). If we compare Table 3.2.98 and the data in appendix A, we find the following: 5 of the 7 consultants who produced the CP sa are members of the highest-ranking Malasi clan (consultants 20, 23, 31, 49, and 51). Consultant 45 is the weather magician heir, and consultant 10 is little Towesei, who was previously mentioned in the discussion of the CPs bukwa and bwa. Thus, all 7 consultants who produced the CP sa are
The System of Classificatory Particles in Kilivila
221
of high status in Tauwema. We may tentatively conclude that the CP sa functions as a sociolinguistic variable. As Table 3.2.215 documents, there were 7 tokens of the CP type sam 'ginger' produced in connection with play accompanying verses. My data do not specify which consultants produced these CP tokens. Thus, I can only note that this CP type is extremely rare in actual speech and that its frame of reference is very distinct: the CP sam is only used to refer to 'ginger' (neya) in play accompanying verses. In other contexts, the CP tarn 'sprouting' is used to refer to ginger. Tables 3.2.99 and 3.2.188 document the production of the CP si 'small bit', which is extremely rare. As Table 3.2.188 documents, 14 of the 15 tokens of this CP type were produced instead of the expected CP ya 'flexible, thin'. Although I explained this use in my comments on Table 3.2.188, it should be noted that the frame of reference of this CP type is not distinct at all. The CP si has obviously been superseded by the CP utu 'scrap'. Moreover, the CPs ke 'wood, inanimates', bwa 'tree, wood', vili 'untwisted', and kwe 'thing' were produced instead of the expected CP si. What is rather astounding is that all 7 consultants who produced this CP type belong to the first three age groups, and that 6 of these consultants are of high status in Tauwema (see appendix A; consultants 4, 25, 26, 31, 10, and 13). However, the database is too small to justify any speculation about the sociolinguistic status of the CP si. Tables 3.2.100 and 3.2.189 document the production of the CP sipu 'sheaf, which belongs to the core of the Kilivila CP system. The frame of reference of this CP is distinct. As Table 3.2.189 shows, the CP sipu is in danger of being superseded by the CP kwe 'thing'. A few tokens of the CPs ya 'flexible, thin' and ke 'inanimates' were produced instead of the expected CP sipw. Tables 3.2.101 and 3.2.190 document the production of the CP sisi 'bough', which belongs to the core of the Kilivila CP system. As Table 3.2.190 documents, the CP sisi has almost superseded the CP type ya 'flexible, thin', at least in its prototypical meaning (see the later discussion of the CP ya). Only 2 tokens of the CP sisi were produced as a response to the stimulus presented to elicit the CP kai 'stone blade'; this use of the CP sisi is not acceptable (see section 3.2.1.2). A few tokens of the CPs ke 'wood, inanimates' and ya 'flexible, thin' were also produced instead of the expected CP sisi. The frame of reference of this CP type is distinct (see also Tables 3.2.216 and 3.2.217). Tables 3.2.102 and 3.2.191 and appendix A document that 62-year-old Weyei produced only 3 tokens of the CP type siva 'time'. This CP type is obviously almost obsolete. It has been superseded by the CP kwe 'thing'. As Table 3.2.191 also records, we find a few tokens of the CPs tuta 'time', utu 'scrap, part', and pila 'part' produced instead of the expected CP siva. As mentioned at the end of section 3.2.2.2, the CP type siwa 'sea portions' was neither produced by my consultants during the elicitation test nor found in the recorded and transcribed texts (see Tables 3.2.216 and 3.2.217). This type has obviously been superseded by the CP kwe 'thing', although my
222
Classificatory Particles in Kilivila
consultants claimed during the informal pretests that this CP is well known and used in their speech community. A few tokens of the repeater bwalita 'sea' and the CP utu 'scrap' were produced instead of the expected CP siwa; however, the use of the CP utu in this context is rather dubious (see section 3.2.1.3). Tables 3.2.103 and 3.2.192 document the production of the CP type suya 'batch of fish'. Although I did not expect this CP, it belongs to the core of the Kilivila CP system. As Table 3.2.192 records, the CP suya has almost superseded the synonymous CP wela 'batch of fish', which was expected as a response to the eliciting stimulus presented. The frame of reference of the CP suya is distinct. Tables 3.2.105 and 3.2.194 document the production of the CP tam 'sprouting (yams)'. The production of this CP, which plays a secondary role in actual speech, was not expected. Its production instead of the expected CP types gula 'heap' and uva 'span' probably resulted from the consultants focusing on the nouns mentioned in the two eliciting noun phrases and not on the general concept expressed by the noun phrases. The production of the CP tarn instead of the expected CP 0 can be explained as follows: the consultants chose to select a more specifying CP type as a more sophisticated response to the stimulus presented to elicit the CP type 0, and a number of them decided to use the CP tarn. Keeping this interpretation in mind, it should be noted that the frame of reference of this CP is distinct (see also Tables 3.2.216 and 3.2.217). Tables 3.2.106 and 3.2.195 and appendix A document that 30-year-old Bwetadou (consultant 45), the weather magician heir produced 1 token of the CP type tetu 'yams'. The production of this repeater as a response to the stimulus presented to elicit the CP 0 was not expected. Its frame of reference is distinct, of course. However, we must note that this CP type is extremely rare and almost obsolete. Tables 3.2.107 and 3.2.196 document the production of the CP to/te 'male persons, human beings', one of the core types of the Kilivila CP system. The variant to is used in the word formation of adjectives and demonstratives, and the varient te is used in the word formation of numerals and the interrogative pronoun/adverb. The frame of reference of this CP is distinct (see also Tables 3.2.216 and 3.2.217). A few tokens of the CP to/te were also produced instead of the expected CPs deli 'group' and yuma 'hand'. The use of the CP to/te as a response to the stimulus presented to elicit the CP yuma is somewhat strange; however, as noted in connection with Table 3.2.196, if we assume that the consultants took the presented stimulus yamala 'her/his hand' as a pars pro to/te to refer to human beings, this use of the CP to/te can be justified. Tables 3.2.108 and 3.2.197 document the production of the CP tuta 'time'. This extremely rare repeater was not expected. Five of my consultants produced a few tokens of this CP instead of the expected CPs siva 'time' and kala 'day'. The frame of reference of this repeater is distinct, of course. Tables 3.2.109 and 3.2.198 document the production of the CP utu 'scrap,
The System of Classificatory Particles in Kilivila
223
parts cut off, which belongs to the core of the Kilivila CP system. Both tables record the broad scope of the frame of reference of this CP type, which has superseded the CP si 'bit' and may have started to supersede the CPs vili 'untwisted' and kabulo1 'point' (see also Tables 3.2.216 and 3.2.217). Table 3.2.198 also records a few tokens of this CP produced by the consultants instead of the expected CPs kabisi 'compartment of a foodhouse', siva 'time', gum 'bit', kabulo2 'sector', meila 'part of a song', nina 'part of a song', nunu 'corner of a garden' bubwa 'parts cut off, and pila 'part'. Moreover, a few tokens of the CP utu were also produced by a few consultants instead of the expected CPs mmwa 'conical bundle', gili 'row', kavi 'tool', kubila 'land plot', and siwa 'sea portions'. The use of the CP utu in these contexts, however, is either not acceptable or rather dubious (see sections 3.2.1.2, 3.2.1.3, and 3.2.1.4). In connection with Table 3.2.198, it should also be noted that the consultants produced some tokens of the CP types ke 'wood, inanimates' and bwa 'tree, wood', as well as a few tokens of the CPs bubwa 'parts cut off, luva 'wooden dishes, tied bundle', and the formative *boma instead of the expected CP utu. As noted in Table 3.2.128, all my other consultants told me that there is no CP *boma in Kilivila. The production of the CP luva as a response to the stimulus presented to elicit the CP utu is not acceptable (see section 3.2.1.5). Tables 3.2.110 and 3.2.199 document the production of the CP uva 'span', which plays a secondary role in actual speech. The frame of reference of this CP is distinct (see also Tables 3.2.216 and 3.2.217). As already mentioned in connection with the CP ke 'long objects, inanimates', the number of tokens of ke that were produced instead of the expected CP uva was obviously a result of the consultants' selective perception of the eliciting noun phrase. This also holds for the production of a few tokens of the CP tam 'sprouting (yams)' instead of uva. Moreover, a few tokens of the CP kwe 'thing' were produced instead of the expected CP uva. The 1 token of the CP na1 documents an unacceptable use of this CP as a response to the stimulus presented to elicit the CP uva. If we compare Table 3.2.110 and the data in appendix A, we find the following: 7 of the 14 consultants producing tokens of this CP are members of the highest-ranking Malasi clan (consultants 31, 38, 41, 49, 52, 59, and 60); 3 consultants belong to the chiefs family (consultants 32, 48, and 53), and consultant 45 is the weather magician heir. Thus, 11 of the 14 consultants producing the CP type uva are of high status in Tauwema. We may infer from this observation that the CP uva functions as a sociolinguistic variable. Tables 3.2.111, 3.2.113, 3.2.200, and 3.2.202 document the production of the CP va/vaya/vayo 'door'. The production of this CP, which is extremely rare, was not expected. Five female consultants produced 14 tokens of the variant va and 2 male consultants produced 3 tokens of the variants vaya/vayo instead of the expected CP duya/kaduyo 'door, entrance'. If we compare Tables 3.2.111 and 3.2.113 and appendix A, we find the following: 4 of the 5 female consultants producing this CP are members of the highest-ranking Malasi clan (consultants 4, 15, 16, and 41), 1 of the 2 male consultants who
224
Classificatory Particles in Kilivila
produced this CP is a member of this clan, too (consultant 34), and the other consultant (45) is the weather magician heir. Thus, 6 of the 7 consultants who produced the CP va/vaya/vayo are of high status. Although the database is rather small, we may tentatively assume that this CP functions as a sociolinguistic variable (especially for women of rank). Tables 3.2.112 and 3.2.201 and appendix A document that the consultants Kapatu (consultant 27) and Bwetadou (consultant 45) produced 1 token each of the obviously extremely rare or almost obsolete CP vakala 'belt' instead of the expected CP gili 'row'. The production of this repeater was not expected. Tables 3.2.114 and 3.2.203 document the production of the CP vili 'untwisted', which plays a secondary role in actual speech. Although a few tokens of the CP vili were produced instead of the expected CPs si 'bit', pila 'part', and gili 'row' (a rather idiosyncratic if not dubious use of the CP vili, indeed; see section 3.2.1.4), the frame of reference of this CP is distinct. In connection with Table 3.2.203, we also note that the consultants produced a number of tokens of the CPs ke 'inanimates' and utu 'scrap', as well as a few tokens of the CPs kwe 'thing', bwa 'tree, wood', and bubwa 'parts cut off, instead of the expected CP vili. The production of the CP bubwa as a response to the stimulus presented to elicit the CP vili is a somewhat dubious use of this CP, however (see section 3.2.1.4). As mentioned at the end of section 3.2.2.2, the CP type vilo was not produced by my consultants during the elicitation test and did not appear in the recorded and transcribed texts (see Tables 3.2.216 and 3.2.217). This CP type is obviously superseded by the CP kwe 'thing'. Two tokens of the CP kada 'road' were produced instead of the expected CP vilo; however, the production of the CP kada as a response to the stimulus presented to elicit the CP vilo is not acceptable. Thus, although my consultants claimed during the informal pretests that the CP vilo is well known and used in their speech community, it seems that this CP type is actually extremely rare or almost obsolete. Tables 3.2.115 and 3.2.204 document the production of the CP vosi 'song'. This repeater, the production of which was not expected, plays a secondary role in actual speech. It was produced by 9 consultants instead of the expected CP meila 'part of a song'. If we compare Table 3.2.115 and the data in appendix A, we find the following: 4 of the 6 female consultants who produced the repeater vosi are members of the highest-ranking Malasi clan (consultants 26, 51, 38, and 41); 1 of the 3 male consultants who produced this CP is the weather magician heir (consultant 45). Thus, 5 of the 9 consultants who produced this CP are of high status in Tauwema. Although the database is rather small, we may tentatively assume that this repeater functions as a sociolinguistic variable (especially for women of status). Tables 3.2.116 and 3.2.205 document the production of the CP wela 'batch of fish', which plays a secondary role in actual speech. As shown in Table 3.2.205, the CP wela also takes over the referential function of the CP type oyla 'string, fish on strings' (see the previous discussion of this CP). However, this finding was expected at the beginning of my research on the CP system: my consultants supplied the noun wela 'batch of fish' as the 'prototypical'
The System of Classificatory Particles in Kilivila
225
noun to elicit the CP oyla. Thus, a number of consultants produced some tokens of the CP wela, as a repeater in this case, as a response to the nominal stimulus presented. As mentioned in connection with Table 3.2.205, the CPs suya 'batch of fish' and na2 'animal' have started to supersede the CP wela. However, if we compare Table 3.2.116 and the data in appendix A, we find the following: 6 of the 14 consultants who produced this CP belong to the highest-ranking Malasi clan (consultants 25, 31, 38, 41, 55, and 60), 3 consultants belong to the chiefs family (consultants 13, 48, and 50), and consultants 45 is the weather magician heir. Thus, 10 of 14 consultants who produced the CP wela are of high status in Tauwema. On the basis of this observation we may infer that the CP wela functions as a sociolinguistic variable. Tables 3.2.117 and 3.2.206 document the production of the CP ya 'flexible, thin', which belongs to the core of the Kilivila CP system. Both tables show that the frame of reference of this CP is rather broad. Although the consultants produced 109 tokens of the CP sisi 'bough' and a few tokens of the CPs ke 'long objects, inanimates', si 'bit', and pila 'part' as a response to the stimulus presented to elicit the CP ya, the CP ya seems itself to have started to supersede the CP gili 'row' and pila 'part', especially in the speech of the consultants in age group II. The CP ya is also rather frequently produced instead of the expected CP duya/kaduyo 'door, entrance'. Moreover, we also find tokens of the CP ya produced instead of the expected CPs bukwa 'fruit cluster', lilo 'walk' (see comments on Table 3.2.206), sipu 'sheaf, sisi 'bough', bubu 'cut across', gum 'bit', and ta.'basket'. One token of the CP ya was also produced instead of the expected CP yuma 'hand'; however, this use of ya is somewhat dubious (see section 3.2.1.4). Thus, we should emphasize that the frame of reference of this CP is not only broad, but also rather indefinite (see also Tables 3.2.216 and 3.2.217). Tables 3.2.118 and 3.2.207 document the production of the CP yam 'day'; this repeater plays a secondary role in actual speech. The CP kwe 'thing' has obviously superseded the CP yam, some tokens of which were also produced by the consultants when the synonymous CP type kala 'day' was expected. If we compare Table 3.2.118 and the data in appendix A, we find the following: 4 of the 7 consultants who produced this CP are members of the highest-ranking Malasi clan (consultants 24, 26, 31 and 38), consultant 21 is the chiefs grandson, and consultant 45 is the weather magician heir. Thus, we record that 6 of the 7 consultants who produced the CP yam are of high status. Although the database is rather small, I tentatively assume that the CP yam functions as a sociolinguistic variable. Tables 3.2.119 and 3.2.208 document the production of the CP yam/yama/yuma 'hand, measure, yard'. This CP is extremely rare or almost obsolete (see also Tables 3.2.216 and 3.2.217). It has almost been superseded by the CPs kweya 'limb' and kwe 'thing'. In connection with Table 3.2.208, we also note the production of a few tokens of the CPs to/te 'human beings', ya 'flexible', and kova 'fire' instead of the expected CP yam/yama/yuma; however, the production of these CPs as a response to the stimulus presented to elicit the CP yam/yama/yuma is a bit strange (see the discussion of the
226
Classificatory Particles in Kilivila
CP to/te above), or dubious (see section 3.2.1.5), or not acceptable (see section 3.2.1.1). However, if we compare Table 3.2.119 and the data in appendix A, we find the following: 3 of the 6 consultants who produced the CP yam/yama/yuma are members of the highest-ranking Malasi clan (consultants 5, 51, and 55), the other 2 consultants are the chiefs grandson (consultant 21) and the chief's wife (consultant 53). Moreover, Kalavatu (consultant IV3) also produced 1 token of this CP in the recorded and transcribed texts (see Tables 3.2.216 and 3.2.217 and appendix B). He is one of the misinari, that is, the local village priests. Although the database is very small, these data indicate that the CP yam/yama/yuma may function as a sociolinguistic variable (especially for women of status). In his description of how to make lime that is eaten together with betel nuts, 41-year-old Tosulala (consultant V17; see appendix B) produced 1 token of the CP type yeni 'a handful of something' as part of a numeral (for body-centered measures, see de Leon 1988:82; Vater 1991:43). This CP type is obviously extremely rare, and I assume that it belongs to the technical language variety of Trobriand lime burners and lime sellers. Tables 3.2.121 and 3.2.210 document the production of the CP yulai/yule 'bundle of four things'. This CP was only produced by 3 male consultants in the elicitation test. In the recorded and transcribed texts (see Table 3.2.216 and 3.2.217 and appendix B), we find 1 token of this CP, also produced by a man (consultant IV1/43). Obviously, this CP is either extremely rare or almost obsolete. Two of the 4 men who produced this CP are members of the highest-ranking Malasi clan (consultants 31 and 55), and consultant IV1/43 is the chiefs son. However, the database is too small to speculate about whether this CP is a sociolinguistic variable, especially for men of rank and high status. As mentioned in connection with Table 3.2.210, the CP kwe 'thing' has obviously superseded the CP yulai/yule. For the sake of completeness I should mention that 2 tokens of the CP ke 'wood, inanimates' were also produced instead of the expected CP yulai/yule, the frame of reference of which, however, is distinct. Tables 3.2.122 and 3.2.211 document the production of the CP yuva 'shoal'. Only 2 consultants, members of the highest-ranking Malasi clan (consultants 30 and 55; see appendix A), produced 4 tokens of this extremely rare and almost obsolete CP. The CP buda/boda 'group, team' seems to have taken over the referential function of the CP yuva. The fact that the CP yuva is almost obsolete is confirmed by the observation that the consultants gave me the "prototypical" noun buda to elicit the CPs yuva and buda (see also the previous discussion of the CP buda/boda). However, the CP kwe 'thing' also is starting to supersede the CP yuva. Moreover, a few tokens of the CPs ke 'inanimates' (general CP), kasa 'row', and kudu1 'band of fibers' were produced instead of the expected CP yuva', however, the use of kasa in this context is rather dubious (see section 3.2.1.2), and the use of kudul in this context is not acceptable (see section 3.2.1.5). Tables 3.2.123 and 3.2.212 document that only 2 consultants, both of high status in Tauwema (see appendix A), produced 1 token each of the CP type
The System of Classificatory Particles in Kilivila
227
0 'basketful of yams', the "zero classifier". In my recorded and transcribed texts, only 2 tokens of this CP type were produced. Although we record 34 tokens of the CP tarn 'sprouting (yams)' and 1 token of the repeater tetu 'yams', the 117 tokens of the CP kwe 'thing' that were produced instead of the expected CP 0 indicate that kwe has almost superseded the CP 0, at least in everyday speech; the CP 0 is only realized during the dadodiga, the festive filling of the yam houses, when baskets of yams are counted. Thus, it has clearly defined context restrictions. 3.3.3.3 Why Are the CP Types Affected by Processes of Language Change, and Who Opposes and Who Fosters Language Change? So far, I have described how the consultants use the various CP types. This description included observations about processes of language change in progress that affect individual CP types. I also inferred from my data the direction of the language change with respect to the CP types. The discussion of the individual CP types illustrated the important role that the frame of reference plays in the production of individual CP types. If we examine these frames of reference now in a summarizing survey, we find the following: 1. If the frame of reference of a certain type loses its distinctness and becomes "fuzzy", and if the meaning of this CP type is not broadened at the same time, then this CP type will be affected by further processes of linguistic change, which may result in this CP type being superseded by another CP type. This observation holds for the CP types bubu/bobu, bubwa, luva, si, and also, to a certain degree at least, for the CP types bwa and gum (see section 3.3.3.2). 2. If the frame of reference of a certain CP type loses its distinctness and even becomes fuzzy, and if the meaning of this CP type is broadened at the same time, then this CP type will affect and may finally even supersede a number of other CP types and their frames of reference. This observation holds for the CP types ke and kwe, of course, but also for the CP types pila, utu, ya, and, to a certain extent, for the CP type sisi and maybe even for the CP types kada, buda/boda/budu, kweya/kwaya/keya, luba, and suya (see section 3.3.3.2). 3. CP types with a rather special, specific, and even distinct, frame of reference may be in danger of being superseded by CP types with broader and more general frames of reference. This observation holds for the CPs bubu/bobu, bubwa, gula, kabisi, kabulo/kabulu, kila, kudu1, lila, lilo/lola, luva, mmwa, nunu, oyla, po'ula, sa, vilo, yam/yama/yuma, yulai/yule, yuva, 0 and also to a certain extent for the CPs bwa,6 kubila/kwabila, kwoila/kwela, liku, nutu/notu, and sipu (see section 3.3.3.2). 6 The CP ke 'wood' ( = ke1) is more common than the CP bwa 'tree, wood', but bwa itself is more general than other CP types in this semantic domain.
228
Classificatory Particles in Kilivila
In addition to the frame of reference, another characteristic of CP types seems to play an important role with respect to processes of language change. A closer look at the CPs bogi, bwalita, yegila, kudu2, tuta, vakala, we/a, yam, duya, and megwa (see section 3.3.3.2) shows that CP types that are repeaters seem to be superseded by CP types that are not repeaters. However, this observation does not hold for the CP luba (which superseded the CP kapwa) and the CP variant pwanina (which is still preferred to the variant pona of the CP type pwanina/pona). A third observation that is relevant in attempting to answer the question of who opposes language change is the following: A number of CP types that are affected by the process of language change are not and will probably never be superseded by other CP types (or at least not easily), because they are produced by a certain group of speakers who use these CP types to mark their societal status. These CP types function as sociolinguistic variables. In section 3.3.3.2, we recorded the following CPs that serve this function: bogi, bubwa, bwa, bwalita, duya/kaduya/kaduyo, gili, kabulo/kabulu2, kudu2, kwoya/koya, liku, nunu, uva, wela, and yam and probably beku, kapwa/kapo, lilo/lola, negwa, peta, sa, and si as well. The CP types kila, va/vaya/vayo, and vosi and probably also yam/yama/yuma are sociolinguistic variables that are especially produced by women of high status. The CP type mmwa and probably also the CP type yulai/yule function as sociolinguistic variables that are especially produced by men of high status. Thus, Trobriand Islanders of high status prevent the processes of linguistic change that would lead to a dramatic decrease of CP types in the Kilivila CP inventory. Some individuals that belong to this group of Trobriand Islanders of high status are also the only consultants who produced a few tokens of CP types that are either extremely rare or almost obsolete; thus, they are keeping these CP types "alive" in the speech community's lexicon. For example, consultants 26, 30, 31, 52, 55, and 59 produced tokens of the almost obsolete CPs beku, yuva, 0, guba, kala, meila, nigwa, nina, siva, yule, and kavi. All these consultants are members of the highest-ranking Malasi clan (see appendix A). Towesei (consultant 10), whose father is also a Malasi and who is very conservative with respect to Trobriand cultural heritage, is the only consultant who produced tokens of the CPs bukwa and dumia. Bwetadou (consultant 45), the weather magician heir, produced tokens of the CPs doba, tetu, vakala, and beku. Sibwesa (consultant 53), the chiefs wife, produced one token of the CP 0. Here, I should emphasize that the previous discussion of the individual CP types revealed that members of the chief's family use a rather broad variety of CP types, especially CPs that function as sociolinguistic variables. We may infer from this that Kilagola, the chief of Tauwema, is very much aware of this function of CPs and supports and promotes the use of these CP types within his family (although he is not related to his children and grandchildren in the Trobriand matrilineal society). It should also be mentioned that experts such as the storyteller Mokopei (consultant V14), the wicker workers Kalavatu (consultant V3) and Tokunupei (consultant V15), and the lime burner and lime seller Tosulala (consultant VI7) will pass on
The System of Classificatory Particles in Kilivila
229
the CP types kali, kauya, lipu, and yeni as part of their technical language variety, together with their expert knowledge, to their respective heirs or successors. Finally, it is clear that consultants whose production of CP types during the elicitation test surpassed the average number of different CP types produced by consultants in the same age group also prevent a too dramatic decrease of CP types in the Kilivila CP inventory. If we compare appendix A and Table 3.2.124, we find the following: 1. Payaya (consultant 4), a member of the Malasi clan, and Towesei (consultant 10), who was mentioned previously, produced the highest number of different CP types within their age group (18 and 29 types, respectively). 2. Within age group II, consultants 15 and 17 produced 33 and 32 different CP types, respectively; both girls are members of the highest-ranking Malasi clan. Consultants 19 and 21 produced 34 and 35 different CP types, respectively; both boys are the chief's grandsons. 3. As for age group III, Kapatu (consultant 27) produced 48 different CP types; Gayoboda (consultant 31), a Malasi, produced 47 different CP types; and Kwelava (consultant 35) produced 46 different CP types. Although Kapatu and Kwelava are members of the third-ranking Lukwasisiga clan, a closer look at my ethnographic data reveals that Kapatu's father, Sakau, and Kwelava's father, Yoya, are members of the highest-ranking Malasi clan. It may well be that Sakau and Yoya also promote and support the elaborate use of the CP type inventory in their children's speech. 4. Within age group IV, consultant 41, who produced 47 different CP types, is a member of the highest-ranking Malasi clan, and consultant 45, who produced 56 different CP types, is the weather magician heir of Tauwema. 5. The two female consultants of age group V, Vadomna (consultant 49) and Kadawaya (consultant 51), produced 41 different CP types each. Weyei (consultant 55) produced 44 different CP types. These three consultants are members of the highest-ranking Malasi clan. Thus, the members of the highest-ranking clan, as well as other consultants of societal status and experts like storytellers, wicker workers, and lime burners, oppose the processes of language change that would result in a decrease of the CP type inventory in Kilivila. Who fosters these processes of language change, then? To answer this question we must check four different sources of information: 1. Table 3.2.124 includes consultants whose production of CP types is clearly below average.
230
Classificatory Particles in Kilivila
2. Tables 3.2.214 and 3.2.215 and the corresponding discussion in section 3.2.1 describe the number of dubious or unacceptable cases of CP production and the consultants who produced them. 3. Tables 3.2.5, 3.2.12, 3.2.19, 3.2.26, and 3.2.33 document the cases of multiple classification, that is, those cases where the consultants produced two or three different CP types as a response to the stimulus that was presented to elicit a particular CP type with an adjective, a numeral, and a demonstrative pronoun. 4. The discussion of the individual CP types given in section 3.3.3.2 includes some useful information about the relationship between the age of the consultants and processes of language change. In addition, we can use information about individual consultants and the general structure of the Trobriand society, if necessary. If we now check Table 3.2.124 and appendix A with respect to the question raised, we find surprisingly that in addition to the consultants of age group I, who produced fewer CP types than average because of their age, consultants 23 (age group II), 25 and 29 (age group III), 44 and 47 (age group IV), 52, 56, 57, and 59 (age group V), who all produced fewer CP types than the average in their respective age group, are all members of the highest-ranking Malasi clan. Consultant 32 (age group III), whose production of different CP types is also below average, is the chief's grandson. Thus, only consultants 14 (age group II; see also section 3.2.1.2), 36 (age group III), and 37 (age group IV), whose production of different CP types is also below average, have low status in Tauwema. This observation seems to contradict our previous conclusion about who opposes the processes of language change. However, this finding shows that processes of language change are so strong that they also affect individuals belonging to groups that in general oppose language change. It should be mentioned here that consultants 52 and 59, who produced less CP types than average, are nevertheless among those consultants who produced tokens of almost obsolete CP types, thus keeping these CP types alive in the speech community. If we compare Table 3.2.214 with the discussion in section 3.2.1 about dubious and unacceptable CP use by individual consultants, and if we check these consultants in appendix A, a similarly puzzling picture emerges: 1. All consultants in age group I, who produced 1 dubious and 8 unacceptable cases of CP production, are members of the highestranking Malasi clan. 2. Of the consultants in age group II who produced the 10 dubious and 22 unacceptable cases of CP production, Olopola (consultant 14) is the only one of low status. Consultants 15, 17, 20, and 23 are members of the Malasi clan, and consultants 13 and 22 are the chiefs grandchildren.
The System of Classificatory Particles in Kilivila
231
3. Of the consultants in age group II who produced the 5 dubious and 14 unacceptable cases of CP production, Kapatu (consultant 27) is the only one of low status; however, as mentioned previously, Kapatu's father is a Malasi. Consultants 25, 30, and 31 are also members of this highest-ranking clan, and consultant 32 is the chief's grandson. 4. Of the consultants in age group IV who produced the 14 dubious and 3 unacceptable cases of CP production, 2 (consultants 41 and 44) are members of the Malasi clan, and 2 (consultants 43 and 48) are the chief's sons. Dakevau (consultant 40) and Igogosa (consultant 42) are the only consultants of low status. 5. Of the consultants in age group V who produced the 4 dubious and 7 unacceptable cases of CP production, 2 are members of the Malasi clan (consultants 49 and 59), and 1 (consultant 50) is the chief's daughter. Kasilasila (consultant 58) is the only consultant who does not belong to a group of status. Again, these findings confirm that the processes of language change are so strong that they also affect individuals belonging to groups that in general oppose such processes of language change. Table 3.2.215 lists the cases of dubious and unacceptable CP type production for all five age groups and differentiates them according to the gender of the speaker. The following picture emerges: 1. The female consultants in age groups I and III produced more dubious and unacceptable cases of CP production than the male consultants within these groups. 2. The female and male consultants in age group II show no differences with respect to the number of cases of dubious and unacceptable CP production. 3. The female consultants in age group IV produced only 3 more cases of dubious and unacceptable CP use than their male peers. 4. The male consultants of age group V produced only 3 more cases of dubious and unacceptable CP use than their female peers. All in all, Table 3.2.215 records that female consultants produced more cases of dubious and unacceptable CP use than male consultants; however, it should also be noted that the consultants of age group II produced most cases of dubious and unacceptable CP use. I interpret these findings as follows: 1. The consultants who belong to groups that oppose linguistic change must be aware of these linguistic processes to some degree. They must also be aware that these processes of language change affect their own speech. Thus, the clash between their "linguistic awareness' (see Sinclair
232
Classificatory Particles in Kilivila
et al. 1978; Bohme 1983:47-70), which opposes these changes, on the one hand, and their own involvement in these processes of language change, on the other hand, may result in "linguistic insecurity" that gives rise to the dubious and even unacceptable cases of CP production recorded in just these consultants' speech. 2. That the girls in age group I are more affected by linguistic insecurity than the boys is easy to understand if we remember that girls in general seem to have more difficulties in acquiring the CP system than boys. 3. That boys and girls of age group II are most seriously affected by linguistic insecurity is also evident: This age group covers the developmental phase that is critical for CP acquisition. In section 3.3.2.1 we noted a kind of explosion in the children's production of CP types, which stops with the consultants in age group III. Thus, this critical phase of the acquisition process documents a strong motivation in these children to use the CP system in an adequate way. This motivation implies a rather high degree of linguistic awareness not only in boys, but also in girls, especially if they are of high status. With the consultants in this age group, we must keep in mind, too, that the 32 cases of dubious and unacceptable CP production are due not only to the clash between their linguistic awareness and the processes of language change in which they themselves are involved, but also to these consultants' readiness and maybe even disposition for a playful exploration of the possibilities such a complex system as the Kilivila CP paradigm offers them. 4. That the male adolescents of age group III clearly show fewer cases of dubious and unacceptable CP production than their female contemporaries, whose behavior is similar to that of the consultants in age group II, is probably due to societal constraints on male adolescents: They must bring their language behavior into line with their claims for societal status. The female adolescents must also obey these societal constraints, convincingly described and analyzed by Weiner (1988), but only after having become adults, that is, generally when they have married, have become mothers, and are thus completely integrated members of the Trobiand adult society (see also Weiner 1976). With respect to the question of who fosters language change, however, we have not yet gathered enough information to offer a convincing answer. Therefore, we will now consider Tables 3.2.5, 3.2.12, 3.2.19, 3.2.26, and 3.2.33 which document the cases of multiple classification. These tables present the number of cases where the stimuli presented to elicit a certain CP type actually elicited two or three different CP types (with the three word classes adjective, numeral, and demonstrative pronoun). I will focus my attention here on the percentages given in the tables; these percentages express the relationship between the number of cases of multiple classification with different elicited CP types and the total number of different CP types
The System of Classificatory Particles in Kilivila
233
Table 3.3.7. The averages of the multiple classification percentages Average m.c. of female consultants, %
Average m.c. of male consultants, %
Average m.c. of all consultants, %
group I
14
8
12
group II
32
28
30
group III
21
12
17
group IV
25
18
21
group V
12
16
14
Consultants, female/male Age 3/3 Age 6/6 Age 6/6 Age 6/6 Age 5/6
m.c. — multiple classification.
produced by all the consultants. Table 3.3.7 summarizes the averages of these multiple classification percentages for the 12 consultants of all five age groups as well as for the female consultants versus the male consultants; it also notes the number of female and male consultants producing each case of multiple classification. Before interpreting this table, I should exlplain why I consider these cases of multiple classification to be relevant to the question raised here. I assume that multiple classification is an indication of at least the following two production strategies that a speaker may pursue: (1) speakers may use multiple classification to display their competence in exploiting the possibilities that the complex CP system offers skillful rhetoricians; (2) speakers may use multiple classification because they are insecure about producing the appropriate CP in a given context. Thus, multiple classification may indicate either skillful language use or linguistic insecurity. Table 3.3.7 shows that with only one exception all consultants in age groups II-V produced cases of multiple classification. The mean or average values presented in the table include, of course, the individual opponents to the observed language changes. We noted previously, however, that the processes of language change are so strong that they affect consultants who belong to groups that oppose these processes of change. Thus, I think it is justifiable to interpret the cases of multiple classification primarily as indications of linguistic insecurity. With this in mind, I will now consider Table 3.3.7 in more detail. The mean values of the consultants in age group II clearly surpass the values for the consultants of the other age groups. The means for the girls of age group II (32%) are slightly higher than for the boys of this age group (28%). The second highest mean values were in age group TV. The consultants in age groups I and V showed only slightly different values if we compare the means for all 12 consultants (12% versus 14%). The consultants in age groups III
234
Classificatory Particles in Kilivila
and V also showed only slightly different values if we compare the means for all 12 consultants (17% versus 14%). If we compare the means for the male and the female consultants, we notice that with the exception of the women of age group V, the values of all other female consultants surpass the values of their male peers. The values for the consultants in age group II and age group IV are of special interest and relevance. Keeping in mind that boys and girls of high status in age group II are most seriously affected by the form of linguistic insecurity that results in the clash between linguistic awareness to oppose changes and involvement in the process of this language change, and keeping in mind the discussion of the CP types bubu/bobu, bwa, deli, kavi, ke, kwe, and ya7 in section 3.3.3.2, which emphasized the role the consultants in age groups II and IV play in the process of language change, especially involving CPs, we can interpret the data presented in Table 3.3.7 in as follows. The data suggest that the language behavior of children in age group II (8-14 years) is not only most seriously affected by the processes of language change within the Kilivila CP system, but also these children could be the agents of change, perhaps because of their previously mentioned readiness and their disposition for playful exploration of the possibilities that the complex CP system offers them. That the children in this age group show much linguistic flexibility and attention, which makes them agents of processes of language change affecting other paradigms in Kilivila, has been documented in connection with Kilivila color terms (G. Senft 1987b). In my research on these color terms, I emphasized that most of my consultants between 8 and 14 years of age were schoolchildren starting to acquire English as a second language (G. Senft 1987b: 338). Six of the schoolchildren who were consultants in my study on color terms also took part in the study on the Kilivila CP system (G. Senft 1987b:321 and appendix A in the present volume). In the sample used in my study of CPs, 8 of the 12 children in age group II attended school regularly; only consultant 16 did not go to school at all. Thus, I propose here, too, that confrontation with a foreign language at school may increase linguistic awareness as well as playful linguistic exploration in these children and perhaps cause them to become agents of processes of language change. The schoolchildren in age group II are joined in their role as agents of the language change by the consultants of age group IV, especially by the women of this age group. I can explain this finding only by relying on my knowledge of the structure of Trobriand society. The consultants in age group IV have acquired the status of adulthood. However, this does not imply that they have also reached the social status that they may strive for within their village community, on the basis of their own personal qualities and possibilities. One way to acquire status in Trobriand society is through linguistic competence, versatility, and elegance (see G. Senft 1985e; 1987d; Weiner 1976; 1983; 1988). To become a skillful and respected speaker. 7
Compare also the tables documenting the production of the CPs na2, sisi, to/te, and utu.
The System of Classificatory Particles in Kilivila
235
Trobrianders of this age are forced to explore the sophisticated and subtle possibilities of their mother tongue. This does not contradict the result presented in section 3.3.2.1, where I stated that the process of acquisition of the CP system can be considered terminated during or after puberty. It is only after the acquisition of this complex system of CPs that speakers can start to use its full semantic power. This kind of revelation and exploration requires readiness and willingness to experiment (see also, e.g., B. Senft and G. Senft 1986:200-221). If speakers realize that a certain level of linguistic competence is reserved to a limited number of persons of high status only, then they may attempt to reach this level of competence if they are of high status themselves, or they may try to reduce these status-making linguistic qualities so that they match their own, more modest, linguistic possibilities. The two groups, being peers, surely influence each other with respect to speech behavior. On the basis of these considerations, it is plausible that consultants in this age group foster (and perhaps are forced to foster) the processes of language change observed within the Kilivila CP system. It is remarkable that this situation is much more evident in the data for female consultants than for male consultants in age group IV (see Table 3.3.7). I can only offer the interpretation that in a matrilineal society women may be more aware that linguistic competence is one expression of the social stratification based on kinship and clan membership. This may also be the key for explaining why we do not note any serious differences in the production of different CP types between men and women in age group IV (see Table 3.2.124 and section 3.3.2.3). On average, women in age group IV, having realized the social importance of the CP system, produce as many different CP types as their male contemporaries; however, as Table 3.3.7 shows, their production of these CP types reveals a higher degree of linguistic insecurity. I can now summarize the findings and interpretations presented in this section with respect to the question of who fosters language change in the CP system: 1. The consultants in age group II foster, and probably also initiate, the processes of language change because of their readiness for a playful exploration of the possibilities the CP system offers and because of their increased linguistic awareness, which may result from their strong motivation to use this system adequately and their confrontation with English as a second language at school. 2. Language change is also fostered and supported by consultants in age group IV, probably because some consultants with low status want to overcome language barriers that mark intrasocietal status, whereas other speakers with high status want to maintain these language barriers. It seems that female consultants in this age group foster language change even more than their male peers, probably because they are more aware of the social stratification of linguistic competence based on kinship and clan membership in their matrilinear society.
236
Classificatory Particles in Kilivila
3.3.3.4 What Are the Possible Consequences of Language Change for the Grammar of Kilivila? As stated at the beginning of the previous section, the frames of reference of the CPs and the characteristics of certain CPs being repeaters or functioning as sociolinguistic variables are crucial qualities for the dynamics of processes of language change that affect the Kilivila CP system. In this section, I will speculate about the direction these processes of language change may take and about the possible consequences of these changes for the grammar of Kilivila. The processes of language change that have been described may result in a rather reduced inventory of core CP types or variants. Although some CP types that function as sociolinguistic variables may be maintained in the lexicon of certain speakers of high status, the use of these CP types may be constrained and constricted to ever more specific contexts. This may, in turn, result in a decrease in the semantic content and power of the CPs. Such a loss must be compensated for in one way or another. As described in chapter 2, the CPs fulfill among other things an important function in constituting Kilivila noun phrases (NPs). Thus, the direction of the changes could lead to a loss of semantic content and information that was previously carried by the CPs in relatively simple NP constructions. The compensation for such a loss may then lead to a more complex expansion of Kilivila NP construction. Thus, more complex attributes or even attributive clauses may take over the functions formerly fulfilled by the complex and highly differentiated CP system. That such a development would also affect the means of securing discourse cohesion, modes of reference, concord, and probably even word formation can be inferred from the description of the functions the CPs fulfill (see chapter 2). In the extreme case, processes of language change affecting the CP system might eventually lead to a complete restructuring of Kilivila grammar. 3.3.4 What Semantic Domains Are Constituted by the Kilivila CP Systems? Now that I have discussed the inventory, the acquisition, and the production of the documented Kilivila CP (sub)system and the processes of language change affecting it, I can proceed to try determine the semantic domains constituted by this system.8 In the first part of this section, I will discuss the semantic domains constituted by the Kilivila CP system described to this point. The second part of this section is an attempt to answer questions about the possible 8 I am aware that the CPs discussed so far represent only a subset of all Kilivila classifiers. I mentioned in section 3.3.1 that Lawton (1980) lists 85 additional CPs, which he collected during his research on Kilivila. However, I do not have references to these CPs in my recorded and transcribed speech data. Thus, I agree with Wittgenstein's maxim: "Wovon man nicht sprechen kann. daruber muB man schweiyen" (Wittgenstein 1971:115).
The System of Classificatory Particles in Kilivila
237
dynamics of these semantic domains; I will try to determine the rules speakers may adhere to in their production of certain CPs. In the final part of this section, I return to the acquisition problems speakers have with the CP system and propose an answer to the question of why the CPs are acquired in the order described in section 3.3.2.4. 3.3.4.1 What Semantic Domains Are Constituted by the Described Subset of the Kilivila CP System? If we examine the English glosses for the CPs given in Table 3.3.1 in section 3.3.1, we can group the 88 CP types into 20 semantic domains that cover the following concepts: 1. a. Persons b. Body parts 2. Animals 3. Quantities a. Living beings in general b. Things in general 4. General CPs (unmarked forms for inanimates) 5. Measures 6. Time 7. Place 8. Qualities 9. Shape 10. Trees, wood, wooden things 11. Utensils 12. Yams 13. Part of a foodhouse, a canoe, a creel 14. Door, entrance, window 15. Fire, oven 16. Road, journey 17. Texts 18. Ritual items 19. Dress, adornment 20. Names The order in which these domains are given here is completely arbitrary. Before I discuss these 20 semantic domains in detail, I want to emphasize that these groupings are primarily based on common sense considerations that, among other things, take into account ethnographic information and knowledge of and about the speech community. One may call such a
Classificatory Particles in Kilivila
238
procedure arbitrary; however, to my mind it is an acceptable and pragmatic (and perhaps the only) starting point and may even be a means for developing a model of the cognitive system that is used to order human perception and is thus the basis for all communication about perceived phenomena. The groupings are based on the fact that all the CPs that constitute a certain semantic domain share certain important features; however, some CPs have inherent features that allow them to be assigned to more than one semantic domain. Such multiple assignment raises questions about the dynamics within such a classificatory system of classifiers. However, before I can discuss such questions, I will present the individual domains and the CPs that constitute them in detail, assuming (or, better, pretending) at this stage of the study that semantic domains can be described as static and more or less closed systems: Domain 1: Persons and body parts The first semantic domain covers the concepts 'person' and 'body part'. I differentiate between these concepts and subdivide the domain into Domain la and Domain 1b. Domain la: Persons The following CP types constitute this subdomain: to/te na
gudi
persons of male gender, human beings' (see Table 3.3.1) 'persons of female gender' ( = na1) (for the meaning 'animals', see Domain 2; for the meaning 'carvings in human likeness', see Domain 10; for the meaning 'stars, moon, planets', see Domain 7; for the meanings 'corpses' and 'spirits, dwarfs', see Table 3.3.1) 'child' (see Table 3.3.1)
Domain 1b: Body parts The following CP types constitute this subdomain: kweya/kwaya/keya yam/yuma/yama ka'i kudu
'limb, severed limb' (see Table 3.3.1) 'hand' (for the meaning 'length', see Domain 5) 'tooth' 'tooth' ( = kudu2) (for the connotation 'band of fibers', see Domain 19).
Domain 2: Animals The following CP types constitute this domain: na kwe
'animals' (but not shells and clams) ( = na2); see also Domain 1 'shells and clams' ( = kwe2); see also Domain 4 and Table 3.3.1)
Domain 3: Quantities Here I differentiate between quantities that generally refer to animate referents (Domain 3a) and quantities that generally refer to inanimate referents (Domain 3b).
The System of Classificatory Particles in Kilivila
239
Domain 3a: Quantities (living beings in general) The following CP types constitute this subdomain: buda/boda/budu deli yuva wela suya oyla
'group, team, crowd' 'company, group on the move' 'shoal' 'batch of fish, string of fish' 'batch of fish on strings' 'fish on strings' (the meaning of this CP has changed from 'string' [see Domain 11] to 'fish on strings').
Domain 3b: Quantities (things in general) The following CP types constitute this subdomain: gula po'ula bukwa duli guba kapwa/kapo luba si gum pila utu kila luva mmwa sa sipu yulai/yule yeni
'heap, group' 'plantation, grove' (see Table 3.3.1) 'fruit cluster' (see Table 3.3.1) 'cluster, bundle' 'bundles of taro' 'bundle (wrapped up), parcel' (see Table 3.3.1) 'bundle (of rolls), parcels (of taro pudding)' 'small bit' 'bit, small piece' 'part, piece' 'scrap, parts (cut off), small particles, fragments' 'clusters of bananas, hands of bananas' 'wooden dishes (full of one's share of food during a food distribution ceremony), tied bundle' 'conical bundle' 'nut bunch' 'sheaf (see Table 3.3.1) 'bundle of four things' 'a handful of something' (see Table 3.3.1)
Domain 4: General CPs The fourth semantic domain covers the "unmarked" forms of inanimates in general. The following CP types constitute this domain: kwe ke
'thing, anything indefinite or unknown' (unmarked form for inanimates in general) ( = kwel) (see Domain 2) unmarked form for inanimates in general ( = ke3); this meaning of ke was not elicited in the test; however, as discussed in section 3.3.3.2, the results of the CP elicitation test show that ke has the status of a general CP, too; for the meaning 'rigid long objects' ( = ke2), see Domain 9; for the meaning 'fire' ( = ke4), see Domain 15; for the meaning 'tree, wooden things' ( = ke1), see Domain 10)
240
Classificatory Particles in Kilivila
Domain 5: Measures The following CP types constitute this domain: uva yuma
'span, measure' (the span of two extended arms from tip to tip; see Table 3.3.1) 'length, measure' (the span of two extended arms from the fingertips of one hand to the wrist of the other hand; see Table 3.3.1)
Domain 6: Time The following CP types constitute this domain: bogi kola siva tuta yam
'night' 'day' 'time' (see Table 3.3.1) 'time, occasion' 'day' (see Table 3.3.1)
Domain 7: Place The following CP types constitute this domain: kabulo/kabulu vilo bwalita siwa kubila/kwabila nunu koya/kwoya dumia na
'village sectors, areas of authority' ( = kabulo2/kabulu; see Table 3.3.1; see also Domain 14) 'place, area, village' 'sea' 'sea portions, ownership divisions with reference to fishing rights' 'large land plot' 'corner(s) of a garden' 'mountain, hill' 'swamp' (see Table 3.3.1) 'stars, planets, moon' ( = na3; see also Domains la, 2, and 10 and Table 3.3.1)
Domain 8: Qualities The following CP types constitute this domain: bubu/bobu/bobo bubwa giwi pwanina/pona/ponina nigwa ya vili
'cut across, cut transversely, (block) cut off 'cut across, (parts) cut off 'cut' 'punctured, something with a hole in it, hole' 'hole' (see Table 3.3.1) 'flexible things, thin things' 'untwisted'
Domain 9: Shape The following CP types constitute this domain: ke kabulo
'rigid, long objects' ( = ke2; see also Domains 4, 10, and 15) 'protuberances' ( = kabulol\ see also Domain 7)
The System of Classificatory Particles in Kilivila kasa gili nutu/notu
241
'row, line' 'row' 'kneaded things, dot, drop'
Domain 10: Trees, wood, wooden things The following CP types constitute this domain: ke bwa sisi lila na
'wooden thing' ( = kel ; see also Domains 4, 9, and 15) 'tree, wooden thing' 'bough, cut off part of a tree' (see Table 3.3.1) 'bough, branch' (see Table 3.3.1) 'carvings in human likeness' ( = na4; see also Domains 1 a, 2, and 7; Table 3.3.1)
Domain 11: Utensils The following CP types constitute this domain: kavi oyla peta/ta kauya kwoila/kwela/kwai
'tool' 'string' (see also Domain 3a) 'basket' (see Table 3.3.1) 'fish trap, creel' (see Table 3.3.1) 'clay pot' (see Table 3.3.1)
Domain 12: Yams The following CP types constitute this domain: tarn 0 tetu
'sprouting, sprouting yams' 'a basketful of yams' 'yams'
Domain 13: Part of a foodhouse/a canoe/a creel The following CP types constitute this domain: lipu kabisi liku
'compartment of a creel, tier' (see Table 3.3.1) 'compartment of a foodhouse, section/division in a foodhouse' 'compartment of a foodhouse, compartment of a canoe' (see Table 3.3.1)
Domain 14: Door, entrance, window The following CP types constitute this domain: duya/kaduyo va/vaya/vayo
'door, entrance' 'door, window' (see Table 3.3.1)
Domain 15: Fire, oven The following CP types constitute this domain: kova ke kumla
'fire, fireplace' 'fire' ( = ke4; see also Domains 4, 9, and 10) 'earth oven'
Classificatory Particles in Kilivila
242
Domain 16: Road, journey The following CP types constitute this domain: kada/keda bililo lilo kali
'road, track' (see Table 3.1.1) 'trip' (see Table 3.3.1) 'walk, journey' (see Table 3.3.1) 'paddle strike' (see Table 3.3.1)
Domain 17: Texts The following CP types constitute this domain: megwa nina meila vosi
'magic, magical formula' 'part(s) of a song' (see Table 3.3.1) 'part(s) of a song, part(s) of a magical formula' (see Table 3.3.1) 'song, part(s) of a song'
Domain 18: Ritual items The following CP types constitute this domain: beku kai sam
'stone blade' (see Malinowski 1922:358; Weiner 1976:179-183) 'stone blade' 'ginger' (in play accompanying verses); (see G. Senft 1985c:70; B. Senft, G. Senft 1986:142; this CP type was not elicited in the elicitation test; see Table 3.3.1)
Domain 19: Dress, adornment The following CP types constitute this domain: doba kudu vakala
'skirt made of banana leaves, "grass" skirt' 'band of fibers' (especially the band of fibers at the waistband of a grass skirt ( = kudul; see also Domain 1b, see Table 3.3.1) 'belt, belt of spondylus shell discs'
Domain 20: Names The following CP type constitutes this domain: iga/yegila
'name'
3.3.4.2 What Are the Dynamics of the Semantic Domains Constituted by the CPs and What Rules Do Speakers Use in Their Production of a Certain CP? The semantic domains presented in the preceding section are based on the assumption, or rather the pretense, that such domains can be described as static or closed systems. However, some of the cross references given for certain CPs, as well as the actual production of the individual CP types described in section 3.3.3.2, make it obvious that this is an idealization. Such an idealization is necessary to establish the basis for discussing the problem
The System of Classificatory Particles in Kilivila
243
of the dynamics of semantic domains, of course, but it has nothing to do with the actual use of CP types in speech. To tackle the task of describing the semantic domains and the dynamics of their constitution by the CP types in actual speech production more adequately, I propose the following proceeding: First, I use the data presented in Tables 3.2.37-3.2.123 and Tables 3.2.126-3.2.218 and the discussion of the production of the individual CPs in actual speech presented in section 3.3.3.2. On the basis of these data and analyses, I assign to the CP types that constitute the semantic domains described in section 3.3.4.1 the tokens that occurred in elicited and observed speech production. I then note any possible "intradomain substitution" of an expected CP type by tokens of one or more unexpected CP type(s). The CP types involved constitute one and the same semantic domain; hence, I call these observed, but unexpected uses of CP types "intradomain substitution." The notation
indicates tokens of CP are produced instead of the expected tokens of CP
X
Y within one domain.
Next, I note how many tokens of CP types assigned in section 3.3.4.1 to other semantic domains are produced by the consultants instead of tokens of the expected CP type(s) constituting the respective domain examined. Because these CP types and their tokens come from other semantic domains into the respective domain examined, I call these observed, acceptable, but unexpected uses in-domain substitution. In-domain substitution is indicated by the notation + + +. Moreover, I also note how many tokens of CP types that constitute the domain examined are used instead of tokens of other CP types constituting one or more different semantic domain(s). Because these CP types and their tokens, so to speak, "drop out" of the semantic domain examined, I call these observed, acceptable, but unexpected uses of these CP types off-domain substitution. Off-domain substitution is indicated by the notation —. With both in-domain and off-domain substitution, I note the types and the tokens observed in these substitution processes as well as all the domains involved in these processes. To provide a basis for comparison, the relative frequency (r.f.) of the tokens of all CP types is computed and presented in italics. The r.f. is computed on
244
Classificatory Particles in Kilivila
the basis of 180 tokens (60 consultants producing, in the ideal case, three tokens of a certain CP type; note that here the CP 0 'basketful of yams' is again an exception since it is only produced in connection with numerals and thus, for this CP, 60 tokens produced by 60 consultants total 100% of elicitation success). If 180 tokens of a certain CP type were produced in the test, the r.f. is 1.0. If no tokens were produced of a CP type that was expected, the CP type gets the r.f. 0 (see, also, G. Senft 1982:84-86, 203-204). I want to emphasize that these relative frequencies are used for the purposes of comparison only. Naturally, these frequencies are only computed for tokens of CP types elicited in the test. Tokens of CP types that were not elicited in the test but were found in the corpus of the transcribed Kilivila speech data are presented in angle brackets ( ) when I describe the relevant semantic domain. The results of this proceeding are presented in tables that attempt to reflect the dynamics involved in the constitution of all 20 semantic domains by the CP types. After the presentation of these 20 tables, all tokens of the CP types that actually constitute a specific semantic domain (and this includes all tokens of CP types that are noted as in-domain substitution cases) are counted and ordered according to the frequency of their production.9 These figures are then totaled. On the basis of this sum, the r.f. for all tokens in each domain, are computed to give a comparative figure for evaluating the intradomain weighting of each CP type. The r.f.s for all CP types constituting a domain add up to 1.0.10 On the basis of these data, I then propose rules that speakers may use in their production of a certain CP of the semantic domain discussed. I regard these rules as the expression of the transformation of a given semantic concept into an appropriate classifier. The rules must cope with both domain-inherent and domain-affecting dynamics. Thus, most, if not all, of the rules proposed and formulated are actually variable rules (for a discussion and further references, see Romaine 1985; G. Senft 1982, 6-8, as well as sections 4.3.3 and 4.5.3 in the present volume). I dispense with a formal notation of these rules in favor of a comprehensive formulation that I hope will be easier to understand. In my opinion, these rules are the only way to meet the expectation discussed in section 3.1, namely, to predict which CPs a speaker will produce to refer to certain nominal concepts. I believe that variable rules can most appropriately describe and record such dynamic processes. We can now examine the semantic domains in detail. Tables 3.4.1.1a and b attempt to present the dynamics of Domain la and provide an evaluation of the CP types that constitute this domain. Table 3.4.1.1a documents that within Domain la there were 3 cases of off-domain substitution, affecting Domain 1b (body part) and Domain 3a 9
It should be clear that we must ignore the off-domain substitution cases here. Off-domain substitution cases observed with one domain are in-domain substitution cases in other domains! 10 If the r.f.s add up to 1.01 or .99, it is not a computing deficiency; it is the result of rounding off the individual values.
The System of Classificatory Particles in Kilivila
245
Table 3.4.1.1 a. Domain la: Persons Off-domain substitution te/to2
180
LOO
nal
173
.96
gudi
142
.79
3
.02
yam/yiima/yama
(Domain 1b)
11
.06
deli
(Domain 3a)
1
.01
deli
(Domain 3a)
Table 3.4.1.Ib. Domain 1a: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain) 1. 2. 3. 4. 5. 6.
te/to2 'male' nal 'female' gudi 'child'
180 173 142
te/to1 'human beings' na5 'corpses' na6 'spirits, dwarfs'
-/- (= not elicited in the test) -/-/-
= 180 = 173 = 142 495
r.f. r.f. r.f.
.36 .35 .29 1. 00
(quantities). Table 3.4.1.1b presents all the CPs that constitute Domain la and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of a certain CP type of Domain la: If speakers want to refer to the concept 'persons of male or female sex' or to the concept 'children', their choice of the CP types te/to2, nal, and gudi is unequivocal (r.f.s 1.0; .96; .79). If speakers want to refer to the concept 'human being', this semantic domain offers the CP to1; if speakers want to refer to the concept 'spirits, dwarfs' or to the concept 'corpses', this semantic domain offers the CP type na in its variants na5 and na6. However, the references to these concepts were not elicited in the test, and thus there are no comparative values to describe their position within this semantic domain. Table 3.4.1.1b documents that the three CPs te/to2, nal, and gudi are of almost equal rank with respect to their importance for the constitution of Domain la, if we base this on language production data similar to those elicited in the test. Tables 3.4.1.2a and b attempt to present the dynamics of Domain 1b and provide an evaluation of the CP types that constitute this domain.
246
Classificatory Particles in Kilivila
Table 3.4.1.2a. Domain 1b: Body part In-domain substitution (Domain 4)
kwe
9
+++
kweya/kwaya/keya
131
0.5 (Domain 4)
ke
5
.73 +++
0.3 (Domain la)
to2
3
+++
yam2/yuma 1/yama 1
0.2 (Domain 4)
kwe
56
12
84-
.07
.47
+++
.31 (Domain 8)
ya
1
+++
.01 ka'i (Domain 4)
ke
2
+++
kudu2
.01 (Domain 4)
kwe
2
26
124
.14
.69
+++
.01
Table 3.4.1.2b. Domain 1b: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
2. 3. 4. 5. 6. 7. 8.
kweya 'limb' ka'i 'tooth' kwe 'thing' kudu2 'tooth' yam/yuma 'hand' ke 'inanimate' te/to2 'male' ya 'flexible'
131 +84 124 9 + 56 + 2 26 12 5+2 3 1
215 124 67 26 12 7 3 1 455
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.47 .27
.15 .06 .03
.02 .01 .002 1.012
Table 3.4.1.2a documents that within Domain 1b there were 2 cases of intradomain substitution and 7 cases of in-domain substitution, affecting CPs that constitute the Domains la (persons), 4 (general CPs), and 8 (qualities). Table 3.4.1.2b presents all the CPs that constitute Domain 1b and gives the r.f.s for the production of the individual CPs within the entire domain. On the basis of these two tables I have formulted the following variable rules that speakers use in their production of a certain CP type of Domain 1b:
The System of Classificatory Particles in Kilivila
247
If speakers want to refer to the concept 'limb (severed or not)', they will generally use the CP kweya/kwaya/keya (r.f. .73); however, it is also possible, although rather rarely observed, to produce one of the two general classifiers (ke (r.f. .03), kwe (r.f. .05)). If speakers want to refer to the concept 'hand', they will only rarely use the special and most appropriate CP type yam/yuma/yama (r.f. .07). As stated in section 3.3.3.2, this CP type functions as a sociolinguistic variable. Instead of this CP type, the CP kweya/kwaya/keya (r.f. .47) will be produced. The production of the general CP kwe is another alternative here (r.f. .31). Moreover, it is also possible, although rather rarely observed, to produce either the CP te/to (r.f. .02) or the CP ya (r.f. .01). If speakers want to refer to the concept 'tooth', they will generally produce the CP ka'i (r.f. .69). Another alternative is the production of the CP kudu2 (r.f. .14); however, this CP variant functions as a sociolinguistic variable (see section 3.3.3.2) and is thus only used by persons of high status. It is also possible, although rather rarely observed, to use one of the general CPs ke, (r.f. .01) or kwe (r.f. .01). As for the cases of in-domain substitution, we may note the following: Table 3.4.1.2b documents that the CPs ya, te/to, ke, yam/yuma/yama, and kudu2 play a secondary role only; the CPs kweya/kwaya/keya, ka'i, and kwe are the most important CPs in Domain 1b, if we base this constitution on language production data similar to those elicited in the test. What should be emphasized here is the rather high value of .75 for the general CP kwe in Table 3.4.1.2b. Obviously, kwe has already invaded a semantic domain the main feature of which is 'animacy'. Thus, the frame of reference of the CP kwe seems to be in the process of broadening from an unmarked form for inanimates to an unmarked form in general. Tables 3.4.2a and b attempt to present the dynamics of Domain 2 and provide an evaluation of the CP types that constitute this domain. Table 3.4.2a. Domain 2: Animals Off-domain substitution
na2 kwe2
162
.90
60
.33
we la
(Domain 3)
53
.29
oyla
(Domain 3 (11))
<19>
Table 3.4.2b. Domain 2: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain) 1. 2
na2 'animals' kwe2 'clams, shells'
162 = 162 -/- (= not elicited in the test)
r.f.
l.(
248
Classificatory Particles in Kilivila
Table 3.4.2a shows that within Domain 2 there were 2 cases of off-domain substitution, affecting Domain 3 (quantities); see Domain 11 (utensils). Table 3.4.2b presents all the CPs that constitute Domain 2 and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of a certain CP type of Domain 2: If speakers want to refer to the concept 'animals' (but not 'shells and clams'), they will only use the CP type na ( = na2 (r.f. .90)). If speakers want to refer to the concept 'shells and clams', this semantic domain offers the CP type kwe (=kwe2). However, the references to these concepts were not elicited in the test, and thus I have no comparative values to describe the position of this CP type within Domain 2 (see also section 3.3.3.2). There were no cases of in-domain substitution and only 2 cases of off-domain substitution; thus, we conclude that Domain 2 is an almost closed
Table 3.4.3.1a Domain 3a: Quantities (living beings in general) In-domain substitution (Domain 4)
kwe
(Domain 9)
kasa
(Domain 4)
ke
(Domain 4)
kwe
(Domain 4)
kwe
(Domain 1)
to
(Domain 1)
nal
(Domain 9)
kasa
Off-domain substitution 85 +++ .47
1 .01 1 .01 52 29 37 .21 11
+++
buda/boda/budu 86 .48
yuva
63 .35
4 .02
+++ +++ +++
deli
93 .52
+++
.06 +++ 1 .01 +++ 2 .01
suya 90 .50
(Domain 2)
(Domain 2)
na2
na2
60
+++
.33 53 +++ .29
wela 30
.17 I1 oyla
12 .07
1 .01 1 .01
mmwa; (Domain 3b) gula; (Domain 3b)
The System of Classificatory Particles in Kilivila
249
Table 3.4.3.1b. Domain 3a: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
kwe 'thing' buda 'group' na2 'animals' deli 'group' suya 'batch/fish' wela 'batch/fish' te/to 'male' yuva 'shoal' nal 'female' kasa 'row' ke 'inanimates'
85 + 37 + 52 63 + 86 60 + 53 93 90 12 + 30 11 4 1 1 +2 1
12.
oyla 'string'/('fish')
0 (= not realized)
= = = = = = = = = = =
174 149 113 93 90 42 11 4 1 3 1 681
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.26 .22 .17 .14 .13 .06 .02 .01 .001 .001 .001 1.013
semantic domain; the frames of reference of its CP types are very distinct indeed. Tables 3.4.3.1a and b attempt to present the dynanics of Domain 3a and provide an evaluation of the CP types that constitute this domain. Table 3.4.3.1a documents that within Domain 3a there were 3 cases of intradomain substitution, affecting CPs that constitute Domains 1 (persons, body parts), 2 (animals), 4 (general CPs), and 9 (shapes). There were also 2 cases of off-domain substitution, affecting Domain 3b. Table 3.4.3.Ib presents all the CPs that constitute Domain 3a and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of a certain CP type of Domain 3a: If speakers want to refer to the concept 'group, team, crowd', they use either the general CP kwe (r.f. .47) or the special and more appropriate CP buda/boda/budu (r.f. .35). The general CP is observed more often than the special classifier; nevertheless, buda/boda/budu belongs to the core of Kilivila CP types (see Table 3.3.6). If speakers want to refer to the concept 'company, group on the move', they most often use the CP deli (r.f. .52); another alternative to refer to this concept is the general CP kwe (r.f. .21). Moreover, it is also possible, although rather rarely observed, to use the CP to/te (r.f. .06) or the CPs nal (r.f. .01) and kasa (r.f. .01) (see Table 3.2.156). If speakers want to refer to the concept 'shoal', they generally produce the CP buda/boda/budu (r.f. .48), thus broadening the frame of reference of this CP from 'group, team, crowd' in such a way that it includes the concept 'shoal'. Another alternative is the general CP kwe (r.f. .29). Moreover, it is also possible, although rarely observed, to refer to this concept either with the special and more appropriate CP yuva (r.f. .02) or the other general CP ke
250
Classificatory Particles in Kilivila
(r.f. .01). To produce the CP kasa in this context seems to be possible; however, it is rather idiosyncratic. If speakers want to refer to the concept 'batch of fish, string of fish', they most often use the CP suya (r.f. .50). Another alternative to refer to this concept is the CP na in its meaning 'animals' (r.f. .33); however, the CP suya is clearly preferred. Only speakers of high status will (rather rarely) produce the special and more appropriate CP wela (r.f. .07) to refer to this concept. As stated in section 3.3.3.2, the CP wela functions as a sociolinguistic variable. If speakers want to refer to the concept 'fish on strings' they use either the CP na in its meaning 'animals' (r.f. .29) or—if they are speakers of high status—the CP wela (r.f. .17). The CP oyla, the special and more appropriate CP to refer to this concept, was not produced in the test and did not occur in my corpus of transcribed Kilivila speech data. As for the cases of in-domain substitution, we may note that the general CP kwe plays the most important role within this domain. Here, the general CP has already invaded a semantic domain that includes 'quantities that generally refer to animate (!) referents'. Although the other general CP (ke) plays only a secondary role within this domain, we can take this finding as another sign for the broadening of the frame of reference of at least the CP kwe to an unmarked form in general. According to Table 3.4.3.1b, the other CPs that play a role within Domain 3a are buda, na2, deli, suya, and—with the restrictions discussed—wela, if we base the constitution of this domain on language production data similar to those elicited in the test. Tables 3.4.3.2a and b attempt to present the dynamics of Domain 3b and provide an evaluation of the CP types that constitute this domain. Table 3.4.3.2a documents that within Domain 3b there were 12 cases of intradomain substitution and 37 cases of in-domain substitution, affecting CPs in Domains 3 a (quantities, living beings in general), 4 (general CPs), 8 (qualities), 9 (shape), 10 (trees, wood, wooden things), and 12 (yams). There were also 19 cases of off-domain substitution, affecting Domains 6 (time), 7 (place), 8 (qualities), 9 (shape), 11 (utensils), 13 (part of a foodhouse/a canoe/a creel), and 17 (texts). Table 3.4.3.2b presents all the CPs that constitute Domain 3b and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of a certain CP type in Domain 3b. If speakers want to refer to the concept 'heap, group '(inanimate)' they use the general CP kwe (r.f. .46) more often than the special and more appropriate CP gula (r.f. .28). It is also possible to use the CP tam (r.f. .16), if the speaker intends to refer especially to a 'heap or group of sprouting yams'. Use of the CP buda (r.f. .01) in this context seems to be possible; however, the use of this CP, which generally refers to animate groups, is rather idiosyncratic here. If speakers want to refer to the concept 'plantation, grove', they use the general CP kwe (r.f. .52) more often than the special and more appropriate
The System of Classificatory Particles in Kilivila
251
Table 3.4.3.2a Domain 3b: Quantities (things in general) Off-domain substitution
In-domain substitution (Domain 3a)
buda
(Domain 12)
tam
(Domain 4)
kwe
1 .01 28 .16 82 .46
+ ++
gula
50 .28
+ ++ +++
3 .02 (Domain 4)
kwe
93 .52
+ ++
po'ula
49 .27 sa below
1
.01 (Domain 10)
ke
(Domain 4)
kwe
(Domain 8)
ya
(Domain 4)
kwe
(Domain 4)
ke
(Domain 4)
kwe
122 .68 4 .02 14 .08
++ +
29 .16 6 .03 30 .17
++ +
bukwa
3 .02
+++
Iuva2 below 2 .01
+++
dull
108 .60
kapwa
46 .26
+++ + ++
sa below 3 .02
77 .43
(Domain 4)
ke
(Domain 4)
kwe
(Domain 10)
bwa
(Domain 8)
vili
28 .16 3 .02 6 .03 4 .02
+++ +++ +++ +++
luba si
1 .01
utu below 118 .66
14 .08
ya; (Domain 8)
Classificatory Particles in Kilivila
252
Table 3.4.3.2a. (continued) Off-domain substitution
In-domain substitution (Domain 4)
kwe
(Domain 4)
ke
(Domain 8)
ya
(Domain 9)
kasa
(Domain 4)
ke
(Domain 4)
kwe
(Domain 8)
ya
(Domain 8)
vili
22 .12 6 .03 6 .05 3 .02
5 .03 3 .02 60 .33 1 .01
+++
gum
88 .49
+++
utu below 1 .01
+++ +++
+++
pila
81 .45
+++ +++ +-H-
91 .57 67 .37 4 .02 2 .01 1
.01 1 .01
1
.01
1
.01 see pila above 1 .01
see gum above see si above (Domain 10)
ke
(Domain 10)
bwa
(DomainS)
bubwa
36 .20 20 .11 8 .04
meila; (Domain 17) nina; (Domain 17) ya; (Domain 8) oyla; (Domain 11) kavi; (Domain 11) siva; (Domain 6) kabisi; (Domain 13) kubila; (Domain 7)
++ + +++ ++ +
utu
87 .48
30 .17 38 .21 4 .02 6 .03 1 .01 4 .02 3 .02 2 .01
vili; (Domain 8) kabulol; (Domain 9) kabisi; (Domain 13) siva; (Domain 6) kabulol; (Domain 7) meila; (Domain 17) nina; (Domain 17) nunu; (Domain 7)
The System of Classificatory Particles in Kilivila
253
Table 3.4.3.2a. (continued) Off-domain substitution
In-domain substitution
3 .03 2 .01 (Domain 4)
ke
(Domain 4)
kwe
142 .79 8
+++
kila
7 .04
++ +
.04
I I see bukwa above
luva2 3 .02
see sa below
guba 3 .02 (Domain 3a)
buda
(Domain 4)
ke
(Domain 4)
kwe
1
+++
121 .67 4 .02
+++
.01
mmwa
+++
luva2 3 .02
po'ula 3 .02 (Domain 10)
ke
(Domain 4)
kwe
(Domain 4)
ke
(Domain 4)
kwe
(Domain 8)
ya
(Domain 4)
ke
(Domain 4)
kwe
114 .63 20 .11 2 .01 49 .27 7
.04 2 .01 148 .82
23 .13
+++
see bukwa above
sa
13 .07
sipu
88 .49
yulai
9 .05
yeni
<\>
+++ +++ +++ +++ +++ +++
siwa; (Domain 7) bubwa; (Domain 8)
Classificatory Particles in Kilivila
254
Table 3.4.3.2b. Domain 3b: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2. 3. 4. 5.
6.
7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.
fce; 'wood' fe5 'inanimates' kel+3 kwe 'inanimates' utu 'scrap' dull 'cluster' gum 'bit' sipu 'sheaf' ya 'flexible' pila 'part' luba 'bundle' gula 'heap' po'ula 'grove' kapwa 'parcel' tarn 'sprouting' bwa 'tree' mmwa 'bundle' sa 'nut-bunch' yulai '4/bundle' luva2 'bundle' bubwa 'cut off kila 'cluster/bananas' vili 'untwisted' bukwa 'cluster' kasa 'row' guba 'bundle/taro' buda 'group' si 'bit'
122+36+114 6 + 28 + 6 + 5 + 142 + 121 + 2 + 2 82 + 93 + 4 + 29 + 30 + 3 + 22 + 3 + 8 + 4 + 20 + 49 + 148 118 + 1 + 1 + 87 108 88 88 14 + 6 + 60 + 7 81 77 50 + 3 49+1+3 46 28 6 + 20 23 3+13 9 2+3+3 8 7 5 3 3 3 1+ 1 1
272
(r.f.
.13)
312 584
(r.f. r.f.
.15) .28
495 207 108 88 88 87 81 77 53 53 46 28 26 23 16 9 8 8 7 5 3 3 3 2 i
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f: r.f. r.f. r.f. r.f.
.24 .10 .05 .04 .04 .04 .04 .04 .03 .03 .02 .01 .01 .01 .01 .004 .004 .004 .003 .002 .001 .001 .001 .001 .001 1.012
2,109 27.
yeni 'a handful' -/- (= not elicited in the test)
CP po'ula (r.f. .27). It is also possible, although rather rarely observed, to use the CP gula (r.f. .02) in this context. If speakers want to refer to the concept 'fruit cluster', they most often use the CP he in its meaning 'wooden things' (r.f. .68), obviously referring to clusters of fruit with wooden shells, skins, or husks like the omnipresent coconuts or the very popular betel nuts. It is also possible to use the CP ya (r.f. .08) in this context, thus emphasizing the flexible quality of fruit clusters. Use of the special and more appropriate CP bukwa (r.f. .02) is as rarely observed as the production of the general CP kwe (r.f. .02) and the CPs sa (r.f. .02), luva (r.f. .01; 'tied bundle of fruit'), and po'ula (r.f. .01; here the speaker may intend to refer to the fruit clusters found on plantations and in groves).
The System of Classificatory Particles in Kilivila
255
If speakers want to refer to the concept 'cluster, bundle', they generally use the special and more appropriate CP dull (r.f. .60). Another alternative is the general CP kwe (r.f. .16). It is also possible to use the general CP ke; however, this is rather rarely observed (r.f. .03). If speakers want to refer to the concept 'bundle (wrapped up)', they most often use the CP luba (r.f. .43; 'bundle of rolls'); obviously, this CP type has broadened its frame of reference. Only speakers of high status produce the sociolinguistic variable kapwa/kapo, the special and most appropriate CP in this context (r.f. .26). It is also possible to use the general CP kwe (r.f. .17) here. If speakers want to refer to the concept 'small bit', they most often use the CP utu (r.f. .66). The special and more appropriate CP si is as rarely produced (r.f. .01) as the CPs vili (r.f. .02; here the speaker may intend to refer to a 'small untwisted bit'), kwe (r.f. .02), and bwa (r.f. .03; here the speaker may intend to refer to a 'small bit of wood'). A more often produced alternative is the general CP ke (r.f. .16). If speakers want to refer to the concept 'bit, small piece', they most often use the special and more appropriate CP gum (r.f. .49). They may also use the general CP kwe (r.f. .12); moreover, it is also possible, although rather rarely observed, to produce the CPs ke (r.f. .03), ya (r.f. .03; here the speaker may intend to emphasize the 'flexibility' or 'thinness' of a 'small piece' of something)', kasa (r.f. .02; here the speaker may intend, somewhat idiosyncratically, to refer to the concept 'bits in a row'), and utu (r.f. .01). If speakers want to refer to the concept 'part, piece', they either use the special and more appropriate CP pila (r.f. .45) or, although less frequently, the CP 'ya' (r.f. .33; here the speaker intend to emphasize the 'thinness' of a 'part' or 'piece' of something). It is also possible, although rather rarely observed, to produce the two general CPs ke (r.f. .03) and kwe (r.f. .02), and the CPs vili (r.f. .01) and utu (r.f. .01). If speakers want to refer to the concept 'scrap, parts (cut off), small particles, fragments', they most often use the special and more appropriate CP utu (r.f. .48). Another alternative is to produce either the CP ke (r.f. .20) or bwa (r.f. .11) to emphasize that 'cut off parts' on the Trobriands are in general 'wooden things'. It is also possible, although rather rarely observed, to use the CP bubwa (r.f. .04) to emphasize the 'cut off' aspect of the concept 'parts cut off'; however, only speakers of high status use this CP, which functions as a sociolinguistic variable (see section 3.3.3.2). If speakers want to refer to the concept 'clusters/hands of bananas', they most often use the general CP ke (r.f. .79). Only women of high status use the special and more appropriate CP kila (r.f. .04); however, the use of this sociolinguistic variable is rather rarely observed. It is also possible, although rarely observed, to use the general CP kwe (r.f. .04) in this context. If speakers want to refer to the concept 'conical bundle', they produce the special and more appropriate CP mmwa only if they are persons of high status (r.f. .13); mmwa functions as a sociolinguistic variable. To refer to this concept, speakers most often use the general CP ke (r.f. .67). It is also possible, although
256
Classificatory Particles in Kilivila
rather rarely observed, to produce the CPs luva (r.f. .02), guba (r.f. .02), kwe (r.f. .02), and buda (r.f. .01); however, the production of the CP buda is somewhat idiosyncratic, if not dubious, in this context. If speakers want to refer to the concept 'nut bunch', they use the special and appropriate CP sa only if they are persons of status (r.f. .13); sa functions as a sociolinguistic variable. To refer to this concept, speakers most often use the CP ke (r.f. .63). It is also possible, although not often observed, to use the CPs kwe (r.f. .11), luva (r.f. .02), andpo'ula (r.f. .02); however, the use of po'ula in this context is rather idiosyncratic. If speakers want to refer to the concept 'sheaf, they most often use the special and appropriate CP sipu (r.f. .49). It is also possible to use either the general CP kwe (r.f. .27) or the CP ya (r.f. .04) or the general CP ke (r.f. .01); however, the production of the last two CPs is rather rarely observed. If speakers want to refer to the concept 'bundles of four things', they usually use the general CP 'kwe' (r.f. .82); the use of the other general CP 'ke' (r.f. .01) is possible, but rather rarely observed. The production of the special and appropriate CP 'yulai' is only rather rarely observed (r.f. .05). If speakers want to refer to the concept 'a handful of something', this semantic domain offers the CP type yeni. However, reference to this concept was not used in the test, and thus there is no comparative value to describe the position of this CP within Domain 3b (see also section 3.3.3.2). As for the cases of in-domain substitution, we may note the following: With the exception of the CPs kwe, ke ( = kel, and ke3), and perhaps ya, all the other CPs (buda, bwa, bubwa, kasa, tam, vili) obviously play only a minor role. As Table 3.4.3.2b shows, Domain 3b is dominated by the CPs kwe, ke, and utu, although there are 27 CPs in this domain. Tables 3.4.4a and b attempt to present the dynamics of Domain 4 and provide an evaluation of the CPs that constitute this domain. Table 3.4.4a documents that within Domain 4 there were 82 cases of off-domain substitution, affecting 17 domains (Domains 3a and 3b are counted as one domain). Table 3.4.4b presents the two CP types that constitute this domain and gives the r.f.s for the production of the CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of the CP types of Domain 4: If speakers want to refer to the concept 'thing, anything unknown or indefinite', they produce the unmarked form, the general CP kwe (r.f. .91). This semantic domain also includes the CP ke in its meaning 'unmarked form (for inanimates)' ( = ke3) as another general CP. However, this CP was not elicited in the test. Domain 4 shows cases of off-domain substitution, but no cases of in-domain substitution or intradomain substitution. Thus, we note that the CPs of Domain 4 play an important role in other semantic domains, but Domain 4 itself is rather closed; it is not affected by the dynamic processes of other domains nor by such processes internally. Tables 3.4.5a and b attempt to present the dynamics of Domain 5 and
The System of Classificatory Particles in Kilivila
257
Table 3.4.4a. Domain 4: General CPs kwe 'thing'
164
.91
56 9 2 85 37 52 82 93 4 29 30 3 22 3 8 4 20 49 148 8 138 144 134 141 40 153 137 53 124 132 28 1 11 23 4 9 29 33 2 6 14 1 49 117 10 55 25 1 1 131 9 61
.31 .05 .01 .47 .21 .29 .46 .52 .02
.16 .17 .02 .12 .02 .04 .02 .11 .27 .82 .04
.77 .80 .74 .78 .22 .85 .76 .29 .69 .73 .16 .01 .06 .13 .02 .05 .16 .18 .01 .03 .08 .01 .27 .65 .06 .31 .07 .01 .01 .73 .05 .34
yam/yama kweya/keya kudu huda/boda deli yuva gula po 'ula bukwa duli kapwa/kapo si gum pila kila mmwa sa sipu yulai/yule uva bogi kala siva yam kabulo2 vilo siwa kubila nunu kwoya/koya bubu/bobu bubwa giwi pwanina vili kasa gili nutu/notu luva kavi oyla peta/ta kwoila 0 kabisi liku duya/kaduyo kova kumla lilo meila nina
(Domain 1) (Domain 1) (Domain 1) (Domain 3a) (Domain 3a) (Domain 3a) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 5) (Domain 6) (Domain 6) (Domain 6) (Domain 6) (Domain 7) (Domain 7) (Domain 7) (Domain 7) (Domain 7) (Domain 7) (Domain 8) (Domain 8) (Domain 8) (Domain 8) (Domain 8) (Domain 9) (Domain 9) (Domain 9) (Domain 10) (Domain 11 ) (Domain 11) (Domain 11) (Domain 11) (Domain 12) (Domain 13) (Domain 13) (Domain 14) (Domain 15) (Domain 15) (Domain 16) (Domain 17) (Domain 17)
258
Classificatory Particles in Kilivila
Table 3.4.4a. (continued) 2 145 6 5 2 1 6 28 6 5 142 149 121 2 2 110 92 1 18 130 16 18 49 2 7 34 5 9 1 12
ke (=ke3) -I-
kai kudu iga/yegila kweya/keya kudu yuva duli si gum pila kila luva mmwa sipu yulai/yule uva kabulo2 kubila bubu/bobu bubwa giwi ya vili kasa kavi oyla kwoila duya/kaduyo lilo meila
.01 .81 .03 .03 .01 .01 .03 .16 .03 .03 .79 .83 .67 .01 .01 .61 .51 .01 .10 .72 .09 .10 .27 .01 .04 .19 .03 .05 .01 .07
(Domain 18) (Domain 19) (Domain 20) (Domain Ib) (Domain Ib) (Domain 3a) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 3b) (Domain 5) (Domain 7) (Domain 7) (Domain 8) (Domain 8) (Domain 8) (Domain 8) (Domain 8) (Domain 9) (Domain 11) (Domain 11) (Domain 11) (Domain 14) (Domain 16) (Domain 17)
Table 3.4.4b. Domain 4: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain) 1.
kwe 'thing'
164
2.
ke 'inanimate'
-/- (= not elicited in the test)
= 164
Table 3.4.5a. Domain 5: Measure (Domain 4)
ke
(Domain 4)
kwe
(Domain 12)
tam
110 +++ .61 8 +++
uva
31 .17
.04
8 +++
.04
yuma
<]>
(r.f.
1.00)
The System of Classificatory Particles in Kilivila
259
Table 3.4.5b. Domain 5: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2. 3. 4.
ke 'inanimates' uva 'span' kwe 'thing' tarn 'sprouting'
5.
yuma'length'
110 31 8 8
110 31 8 8 157
r.f. r.f. r.f. r.f.
.70 .20 .05 .05 1.00
-/- (=not elicited in the test)
provide an evaluation of the CP types that constitute this domain. Table 3.4.5a documents that within Domain 5 there were 3 cases of in-domain substitution, affecting CPs that constitute Domains 4 (general CPs) and 12 (yams). Table 3.4.5b presents all the CPs that constitute Domain 5 and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 5: If speakers want to refer to the concept 'span, measure, the span of two extended arms, from tip to tip', they use the special and appropriate CP uva only if they are persons of high status; the CP uva functions as a sociolinguistic variable (r.f. .17). Usually the general CP ke (r.f. .61) is used in this context. It is also possible, although rather rarely observed, to use the other general CP kwe (r.f. .04) or the CP tarn (r.f. .04; here, the speaker intends to refer to a "kuvi"-type of long yams that are measured in such a way, thus emphasizing the item measured). This domain also includes the CP yuma to refer to the concept "length, measure; the span of two extended arms from the fingertips of one hand to the wrist of the other hand'. However, this CP type was not elicited in the test. It was used only once in my corpus of transcribed Kilivila speech data. Thus, we have no comparative value to describe the position of this CP type within Domain 5 (see also section 3.3.3.2). Table 3.4.5b shows that the two CPs ke ( = ke3) and uva are the most important CPs in Domain 5. I should note here that the CP 0 could also be included as one of the CP types in Domain 5. However, it could also be argued that the zero classifier actually belongs to Domain 3b. I decided to place the zero classifier in Domain 12. This CP illustrates the difficulties inherent in grouping classifiers into semantic domains (see section 3.3.4.1). Tables 3.4.6a and b attempt to present the dynamics of Domain 6 and provide an evaluation of the CP types that constitute this domain. Table 3.4.6a documents that within Domain 6 there were 3 cases of intradomain substitution and 6 cases of in-domain substitution affecting CPs in Domains 3b (quantities) and 4 (general CPs).
Classificatory Particles in Kilivila
260
Table 3.4.6a. Domain 6: Time In-domain substitution (Domain 4)
kwe
(Domain 4)
kwe
(Domain 4)
kwe
138 +++ .77 144 +++ .80 141 +++ .78
bogi kala
yam
12 .07 3 .02 14 .08
5
.03 1 .01
tuta 1 .04
(Domain 4)
kwe
(Domain 3b)
pila
(Domain 3b)
utu
134 +++ .74 1 +++ .01 6 +++ .03
siva
3 .02
meila
Table 3.4.6b. Domain 6: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2. 3. 4. 5. 6. 1. 8.
kwe 'thing' yam 'day' bogi 'night' tuta 'time' utu 'fragment' kala 'day' siva 'time' pila 'part'
9.
meila 'part of.a day' -/- (=not elicited in the test)
138 + 144 + 134 + 141 14 + 5
12 8 6 3 3 1
557 19 12 8 6 3 3 1 609
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.91 .03 .02 .01 .01 .01 .01 .002 1.002
Table 3.4.6b presents the CP types that constitute this domain and gives the r.f.s for the production of the CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 6: If speakers want to refer to the concept 'night', they usually use the general CP kwe (r.f. .77). Only if they are persons of high status, do they use the special and most appropriate CP bogi (r.f. .07); bogi functions as a sociolinguistic variable (see section 3.3.3.2). If speakers want to refer to the concept 'day', they usually use the general CP kwe (r.f. 80). The special and appropriate CP kala is only very rarely used (r.f. .02). Here, it is also possible, although rather rarely observed, to
The System of Classificatory Particles in Kilivila
261
use the CP tuta (r.f. .01), including the more specific concept 'day' into the more general concept 'time'. Moreover, speakers of high status may also use the second special and appropriate CP yam ( = yam1); however, the use of this CP is rather rarely observed (r.f. .03). This holds also for the use of this sociolinguistic variable if it is observed separately: The use of the CP yam—referring to the concept 'day'—is only rather rarely observed (r.f. .08); here, again, most speakers use the general CP kwe (r.f. .78). If speakers want to refer to the concept 'time', they usually use the general CP kwe (r.f. .74). The production of the two special and appropriate CPs tuta (r.f. .04) and siva (r.f. .02) is only rather rarely observed. It is also possible, although rare, to produce the CPs utu (r.f. .03; here, the speaker may intend to refer to 'parts, fragments of time') andpila (r.f. .01; here, the speaker may intend to refer to 'parts, pieces of time') in this context. However, the use of these two CPs is somewhat idiosyncratic. I should note here that this semantic domain also includes the CP type meila to refer to the concept 'part of a day'. However, this meaning of the CP meila was neither elicited in the test, nor observed in my corpus of transcribed Kilivila speech data. I discovered the meaning of meila during my lexicographic work with consultants. Table 3.4.6b shows that the general CP kwe dominates Domain 6. Tables 3.4.7a and b attempt to present the dynamics of Domain 7 and provide an evaluation of the CP types that constitute this domain. Table 3.4.7a documents that within Domain 7 there was only 1 case of intradomain substitution and 13 cases of in-domain substitution, affecting CPs that constitute Domains 3b (quantities), 4 (general CPs), and 9 (shape). Table 3.4.7b presents the CP types that constitute this domain and gives the r.f.s for the production of the CP types within the domain as a whole. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 7: If speakers want to refer to the concept 'village sector, area of authority', they use the special and appropriate CP kabulo2/kabulu (r.f. .12) only if they are persons of high status; the CP kabulo2/kabulu functions as a sociolinguistic variable. To refer to this concept, speakers usually use either the general CP ke (r.f. .51) or, although less frequently, the general CP kwe (r.f. .22). It is also possible, although rather rarely observed, to produce the CP utu (r.f. .01). If speakers want to refer to the concept 'place, area, village', they use the general CP kwe (r.f. .85). The special and appropriate CP vilo seems to have become obsolete; this CP is no longer observed in actual speech. If speakers want to refer to the concept 'sea portions, ownership division with reference to fishing rights', they usually use the general CP kwe (r.f. . 76). It is also possible, although rather rarely observed, to use the CP bwalita (r.f. .03), if the speaker is a person of high status, or to use the CP utu (r.f. .02), although the CP utu in this context is somewhat idiosyncratic. The special and appropriate CP siwa seems to have become obsolete; the realization of this CP is no longer observed in actual speech production. If speakers want to refer to the concept 'large land plot', they usually use
262
Classificatory Particles in Kilivila
Table 3.4.7a. Domain 7: Place In-domain substitution (Domain 4)
ke
(Domain 4)
kwe
(Domain 3b)
utu
(Domain 4)
kwe
92 .51 40 .22 1 .01 153
+++
kabulo2/kabulu
21 .12
+++ +++ +++
.85
vilo bwalita
(Domain 4)
kwe
(Domain 3b)
utu
(Domain 4)
kwe
(Domain 4)
ke
(Domain 3b)
6 .03
137 +++ .76 3 +++
siwa
kubila/kwabila
94 .52
pila
53 +++ .29 1 +++ .01 1 +++
(Domain 4)
kwe
124 +++
nunu
21 .12
(Domain 3b)
utu
(Domain 9)
kabulol
(Domain 4)
kwe
koya/kwoya
26 .14
.02
.01 .69 2 +++
.01
3 +++ .02 132 +++
.73 dumia
<1>
na3
<4>
the special and appropriate CP kubila/kwabila (r.f. .52). The general CP kwe (r.f. .29) is also produced in this context. Moreover, it is also possible, although rather rarely observed, to produce the general CP ke (r.f. .01) or the CP pila (r.f. .01). If speakers want to refer to the concept 'corner(s) of a garden', they use the special and appropriate CP nunu (r.f. .12) only if they are persons of high status. To refer to this concept, the general CP kwe (r.f. .69) is usually used. It is also possible, although rather rarely observed, to use the CP kabulol/kabulu (r.f. .02; here the speaker may intend to refer to the way in which the corners of a garden are marked), or to use the CP utu (r.f. .01) in this context.
The System of Classificatory Particles in Kilivila
263
Table 3.4.7b. Domain 7: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
kwe 'thing'
2. 3. 4. 5. 6. 7. 8. 9. 10.
kubila 'land plot' ke 'inanimates' koya 'mountain' kabulo2 'sector' nunu 'corner' utu 'part' bwalita 'sea' kabulol 'point' pila 'part'
11. 12. 13. 14. 15.
siwa 'sea portion' dumia 'swamp' na3 'stars, moon' liku 'authority area' va/vaya 'river, creek'
40 + 153 + 137 + 53 + 124 + 132 94 92 + 1 26 21 21 1+3+2 6 3 1
639 94 93 26 21 21
6 6 3 1 910
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.70 .10 .10 .03
.02 .02 .01 .01 .003 .001 .994
0 (= not realized) -/- (= not elicited in the test) -/-/-/-
If speakers want to refer to the concept 'mountain, hill', they only use the special and appropriate CP koya/kwoya (r.f. .14) if they are persons of high status. To refer to this concept, the general CP kwe (r.f. .73) is most often produced. Domain 7 also include the CP dumia to refer to the concept 'swamp' and the CP na to refer to the concept 'stars, planets, moon' ( = na3). However, these two CPs were not elicited in the test, and thus we have no comparative values to describe the position of these CPs within Domain 7. Moroever, I should note that Domain 7 also includes the CP liku to refer to the concept 'area of authority' and the CP va/vaya/vayo to refer to the concept 'river, creek, sea passage'. However, the meanings of these CPs were not elicited in the test; they are the result of my lexicographic work. Table 3.4.7b shows that Domain 7 is also dominated by the general CP kwe. Tables 3.4.8a and b attempt to present the dynamics of Domain 8 and provide an evaluation of the CP types that constitute this domain. Table 3.4.8a documents that within Domain 8 there were 3 cases of intradomain substitution and 19 cases of in-domain substitution, affecting CPs in Domains 3b (quantities), 4 (general CPs), 9 (shape), and 10 (trees, wood, wooden things). There were also 14 cases of off-domain substitution, affecting CPs in Domains Ib (body parts), 3b (quantities), 9 (shape), 10 (trees), 11 (utensils), 14 (door, window, entrance), and 16 (journey, road). Table 3.4.8b presents the CP types that constitute Domain 8 and gives the r.f.s for the production of the individual CP types within the entire domain.
Classificatory Particles in Kilivila
264
Table 3.4.8a. Domain 8: Qualities Off-domain substitution
In-domain substitution (Domain 4)
ke
(Domain 4)
kwe
(Domain 9)
kasa
(Domain 4)
kwe
(Domain 4)
ke
(Domain 4)
kwe
(Domain 10)
bwa
(Domain 3b)
utu
(Domain 4)
kwe
(Domain 4)
ke
(Domain 10)
bwa
(Domain 4)
ke
(Domain 3b)
pila
(Domain 3b)
si
(Domain 10)
sisi
(Domain 4)
ke
16 .09 11 .06 3 .02 23 .13
130 .72 1 .01 16 .09 2 .01 28 .16 18 .10 3 .02 18 .10 4 .02 14 .08 109 .61
106 .59
+++ giwi +++ +++ +++ pwanina/pona
nigwa +++ bubwa
140 .78
16 .09
1 .01 8 .04
utu (Domain 3b)
+++ +++ +++
2 .01
+++ bubu/bobu/bobo +++ +++ +++ ya
68 .38
4 .02 12 .07
+++ +++ +++
49 +++ vili .27
68 .38
101 .26 60 .33 14 .08 6 .03 1 .04 1 .01 6 .03 1 .01 45 .12 16 .09 4 .02
gili (Domain 9) pila (Domain 3b) bukwa (Domain 3b) gum (Domain 3b) sipu (Domain 3b) yani/yarna (Domain Ib) sisi (Domain 10) peta/ta (Domain 11) duya/kaduyo (Domain 14) Mo (Domain 16) si (Domain 3b)
The System of Classificatory Particles in Kilivila
265
Table 3.4.8a. (continued) Off-domain substitution
In-domain substitution (Domain 4)
kwe
(Domain 3b)
utu
(Domain 10)
bwa
\ .01 3 .02
4 +++ .02 30 +++ .17 2 +++
pila (Domain 3b) gili (Domain 9)
.01
Table 3.4.8b. Domain 8: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
ke3 'inanimates' pwanina 'hole' sisi 'bough' giwi 'cut' bubu 'cut across' vili 'untwisted' kwe 'thing' utu 'scrap' bwa 'tree' bubwa 'cut off ya 'flexible' si 'small bit' pila 'part' kasa 'row' nigwa 'hole'
18 + 130 + 16 + 18 + 49 140 109 106 68 68 28 + 1 + 11 + 23 + 4 2 + 30 3 + 16 + 2 16 + 2 12 + 4 14 4 3 1
231 140 109 106 68 68 67 32 21 18 16 14 4 3 1
898
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.26 .16 .12 .12 .08 .08 .07 .04 .02 .02 .02 .02 .004
.003 .001 1.018
On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 8: If speakers want to refer to the concept 'cut', they usually use the special and appropriate CP giwi (r.f. .59). It is also possible, although not often observed, to use the general CPs ke (r.f. .09) and kwe (r.f. .06) and the CP kasa (r.f. .02; here, the speaker may intend to refer to 'something cut in rows'); however, the production of the CP kasa in this context is somewhat idiosyncratic. If speakers want to refer to the concept 'punctured, something with a hole in it, hole', they usually use the special and appropriate CP pwanina/pona (r.f. .78). It is also possible to use the general CP kwe (r.f. .13) or the almost synonymous CP nigwa (r.f. .01); however, nigwa is only rather rarely observed. If speakers want to refer to the concept 'cut across, parts cut off', they use the special and appropriate CP bubwa only if they are persons of high status (r.f. .09). To refer to this concept, speakers most often use the general CP ke
266
Classificatory Particles in Kilivila
(r.f. .72). It is also possible, although not often observed, to use the CP bwa (r.f. .09; here, the speaker may intend to refer to 'wooden parts cut off', emphasizing the wooden quality of these parts), the CP utu (r.f. .01), and the general CP kwe (r.f. .01) in this context. If speakers want to refer to the concept 'cut across, cut transversely, block cut off', they usually use the special and appropriate CP bubujbobu/bobo (r.f. .38). The two general CPs kwe (r.f. .16) and ke (r.f. .10) are found less frequently in this context. It is also possible, although rather rarely observed, to use the CPs bwa (r.f. .02; here, the speaker may intend to refer to a 'block of wood cut off'), ya (r.f. .02; here, the speaker may intend to refer to 'flexible things that are cut'), and, if the speaker is a person of high status, the almost synonymous CP bubwa (r.f. .01). If speakers want to refer to the concept 'flexible things, thin things', they only rather rarely use the special and appropriate CP ya (r.f. .07). To refer to this concept, the CP sisi (r.f. .61) is most often used; here, the speakers obviously extend the meaning of the CP sisi. Moreover, it is also possible, although not so often if not rather rarely observed, to produce the general CP ke (r.f. .10), the CP si (r.f. .08; here, the speaker may intend to refer to the concept 'a small bit of a thin, flexible thing'), and the CP pila (r.f. .02; here, the speaker may intend to refer to the concept 'apart of a thin, flexible thing') in this context. If speakers want to refer to the concept 'untwisted', they use the special and appropriate CP vili (r.f. .38) more often than the general CP ke (r.f. .27) and the CP utu (r.f. .17; here the speaker may intend to refer to the concept 'untwisted particles, fragments, parts cut off', emphasing their untwisted quality). It is also possible, although rather rarely observed, to use the CPs kwe (r.f. .02) and bwa, although bwa is somewhat idiosyncratic in this context. Table 3.4.8b shows that the CPs ke3, pwanina/pona, sisi, as well as bubu/bobu/bobo, vili, and kwe are the most important CPs in Domain 8. Tables 3.4.9a and b attempt to present the dynamics of Domain 9 and provide an evaluation of the CP types that constitute this domain. Table 3.4.9a documents that within Domain 9 there were 2 cases of intradomain substitution and 9 cases of in-domain substitution, affecting CPs in Domains 3b (quantities), 4 (general CPs), 8 (qualities), 10 (trees, wood, wooden things), and 19 (dress, adornment). There were also 7 cases of off-domain substitution, affecting Domains 3a and b (quantities), 7 (place), 8 (qualities), and 17 (texts). Table 3.4.9b presents the CP types that constitute Domain 9 and gives the r.f.s for the production of the individual CPs within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 9: If speakers want to refer to the concept 'protuberances', they do not use the special and appropriate CP kabulol; instead, they usually use the CP ke2 (r.f. .59). The CP utu is also used in this context (r.f. .21; here, the speaker may intend to refer to a 'protuberant part'). Moreover, it is possible, although rather rarely observed, to use the CP bwa in this context (r.f. .02; here, the speaker may intend to refer to 'proturberant wooden things').
The System of Classificatory Particles in Kilivila
267
Table 3.4.9a. Domain 9: Shape Off-domain substitution
In-domain substitution ke2
(Domain 10)
bwa
(Domain 3b)
utu
(Domain 4)
ke
(Domain 4)
kwe
(Domain 4) (Domain 8) (Domain 8)
kwe vili ya
(Domain 19)
vakala
(Domain 4)
kwe
3 .02 38 .21 1 .01 9 .05
+++
.107 .59
kabulol
3 .02
nunu (Domain 7)
10 .06 3 .02 3 .02 3 .02 1 .01 2 .01
meila (Domain nina (Domain giwi (Domain gum (Domain yuva (Domain deli (Domain
+++
+++
kasa
139 .77
+++
29
+++
3
+++
.16
gili
3 .02
17) 17) 8) 3b) 3a) 3a)
12 .07
.02 101
+++
.56 2
+++
33
+++
.01
nutu/notu
.18
103 .57
Table 3.4.9b. Domain 9: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2.
3. 4. 5. 6. 7. 8. 9. 10.
kasa 'row, line' ke2 'rigid, long' ke3 'inanimates' notu 'kneaded/dot' ya 'flexible/thin' kwe 'thing' utu 'scrap' gili 'row'
vili 'untwisted' bwa 'tree' vakala 'belt'
139 + 3 107 2 + 109 103 101 9 + 29 + 33 38 12 3 3 2
142 107 2
109 103 101 71 38 12 3 3 2
585
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.24
.18) .003) .19 .18 .17 .12 .06 .02 .005 .005 .003
.99
Classificatory Particles in Kilivila
268
If speakers want to refer to the concept 'row, line', they usually use the special and appropriate CP kasa (r.f. . 77). It is also possible, although rather rarely observed, to use the general CPs kwe (r.f. .05) and ke (r.f. .01) in this context. If speakers want to refer to the concept 'row', they use the special and appropriate CP gili only if they are of high status (r.f. .07). To refer to this concept, a speaker usually uses the CP ya (r.f. .56; here the speaker may intend to emphasize the flexible quality of a row). Another possibility is to use the general CP kwe, of course (r.f. .16). Moreover, it is also possible, although rather rarely observed, to use the CP vili (r.f. .02; here the speaker may intend to refer to a 'row brought into an untwisted line'), the almost synonymous CP kasa (r.f. .02), or the CP vakala (r.f. .01); however, vakala in this context is somewhat idiosyncratic. If speakers want to refer to the concept 'kneaded things, dot, drop', they usually use the special and appropriate CP nutu/notu (r.f. .57). It is also possible to use the general CP kwe (r.f. .18) in this context. Table 3.4.9b shows that the CPs kasa, ke3, nutu/notu, ya, and kwe play the most important roles within Domain 9. Tables 3.4.l0a and b attempt to present the dynamics of Domain 10 and provide an evaluation of the CPs that constitute this domain. Table 3.4.10a documents that within Domain 10 there were 5 cases of intradomain substitution and 2 cases of in-domain substitution, affecting CPs in Domains 4 (general CPs) and 8 (qualities). There were also 13 cases of off-domain substitution, affecting Domains 3b (quantities), 8 (qualities), 9 (shape), and 13 (part of a foodhouse/a canoe). Table 3.4.10b presents the CP types that constitute Domain 10 and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 10: If speakers want to refer to the concept 'wood, wooden things', they usually use the special and appropriate CP kel (r.f. .92). It is also possible, although Table 3.4.10a. Domain 10: Tree, Wood, Wooden Thing In-domain substitution
Off-domain substitution kel
166
.92
2 .01
see sisi, lila luva below
114 .63 122 .68 53 .30 32 .18 36 .20
sa (Domain 3b) bukwa (Domain 3b) kabisi (Domain 13) liku (Domain 13) utu (Domain 3b)
The System of Classificatory Particles in Kilivila
269
Table 3.4.l0a. (continued) Off-domain substitution
In-domain substitution
.155 .86 bwa
20 .11 6 .03 16 .09 2 .01 3 .02 4 .02 3 .02 109 .61
5 .03
kel 8 .04 (Domain 8)
ya
6 .03
sisi
+++ I kel 152 .84
140 .78
utu (Domain 3b) si
(Domain 3b) bubwa (Domain 8) vili (Domain 8) kabulol (Domain 9) kabisi (Domain 13) bubu/bobu (Domain 8)
ya
(Domain 8)
lila kel 149 .83 (Domain 4)
kwe
2 .01
luval
+++
na4
4 .02 <3>
Table 3.4.10b. Domain 10: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
kel 'wood'
166 + 155 + 8 + 152 + 149 + 114
2. 3. 4. 5. 6.
sisi 'bough' bwa 'tree' ya 'flexible' luva 'wooden dish' kwe 'thing'
140 6 4 2
7. 8.
lila 'bough' na4 'carving'
0 (= not realized) -/- (= not elicited in the test)
5+2
744 140
7 6 4 2
903
r.f. r.f. r.f. r.f. r.f. r.f.
.82 .16 .01 .01 .004 .002 1.006
270
Classificatory Particles in Kilivila
rather rarely observed, to use the almost synonymous CP bwa (r.f. .01) in this context. However, bwa functions as a sociolinguistic variable and is thus only used by persons of high status. If speakers want to refer to the concept 'trees, wooden things', they also usually use the CP kel (r.f. .86); the special and appropriate CP bwa is rather rarely produced and only by persons of high status (r.f. .03). If speakers want to refer to the concept 'bough, cut-off part of a tree', they usually use the special and appropriate CP sisi (r.f. .78). It is also possible, although rather rarely observed, to use the CP ya (r.f. .03) in this context. If speakers want to refer to the concept 'bough, branch', they use the CP kel (r.f. .84). The special and appropriate CP lila was not produced in the test; moreover, it did not occur in my corpus of transcribed Kilivila speech data. If speakers want to refer to the concept 'wooden dishes', they usually use the CP kel (r.f. .83). Use of the special and appropriate CP luva (r.f. .02) is only rather rarely observed. It is also possible to refer to this concept with the general CP kwe (r.f. .01); however, the use of kwe in this context is only rather rarely observed. Domain 10 also includes the CP na to refer to the concept 'carvings in human likeness' (=na4). However, this meaning was not elicited in the test, and thus we have no comparative value to describe the position of this CP within Domain 10. Table 3.4.10b shows that Domain 10 is dominated by the CP kel and that the only other CP that plays an important role within this domain is sisi. Tables 3.4.11a and b attempt to present the dynamics of Domain 11 and provide an evaluation of the CP types that constitute this domain. Table 3.4.1 la documents that within Domain 11 there were 10 cases of in-domain substitution, affecting CPs in Domains 3b (quantities), 4 (general CPs), and 8 (qualities). Table 3.4.11b presents the CP types that constitute Domain 11 and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 11: If speakers want to refer to the concept 'tool', they use either the special and appropriate CP kavi (r.f. .02), the general CPs ke (r.f. .04) and kwe (r.f. .03), or the CP pila (r.f. .01). However, the production of all these CPs and thus the reference to this concept is only very rarely observed; moreover, the use of pila in this context is rather idiosyncratic. If speakers want to refer to the concept 'string', they produce either the general CP ke (r.f. .19) or, although less frequently, the general CP kwe (r.f. .08). It seems to be also possible to use the CP pila (r.f. .01) in this context; however, the use of pila in this context is rather idiosyncratic. The special and appropriate CP oyla occurred neither in the test nor in my corpus of transcribed Kilivila speech data (see also Domain 3a). If speakers want to refer to the concept 'basket', they usually use the special and appropriate CP peta/ta (r.f. .88). The variant peta, however, is only used
The System of Classificatory Particles in Kilivila
271
Table 3.4.11 a. Domain 11: Utensils In-domain substitution (Domain 4)
ke
(Domain 4)
kwe
(Domain 3b)
pila
(Domain 4)
ke
(Domain 4)
lave
(Domain 3b)
pila
(Domain 4)
kwe
(Domain 8)
ya
(Domain 4)
kwe
(Domain 4)
ke
1 .04 6 .03 1 .01 34 .19 14 .08 2 .01 1 .01 1 .01 49 .27 5 .03
+++
kavi
3 .02
+++ +++ +++
oyla (see also Domain 3a)
+++ +++ +++
peta/ta
158 .88
+++ +++
kwoila
(= .04 + .84)
96 .53
+++
kauya
<8>
Table 3.4.1 Ib. Domain 11: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
peta/ta 'basket'
158
158
r.f.
2.
kwoila 'clay pot'
96
96
r.f.
3.
kwe 'thing'
6+14+1+49
70
r.f.
4.
ke3 'inanimate'
7+34+5
46
r.f.
5.
kavi 'tool'
3
=3
r.f.
6.
pila 'part'
1+2
3
r.f.
7.
ya 'thin'
1
1
r.f.
8.
oylo 'string'
9.
kauya 'utensil'
0 -/-
377
.42 .25 .19 .12 .01 .01 .003 1.003
(= not realized) (= not elicited in the test)
by persons of high status (r.f. .04). It is also possible, although rather rarely observed, to use the general CP kwe (r.f. .01) or the CP ya (r.f. .01). If speakers want to refer to the concept 'clay pot, potlike', they usually use the special and appropriate CP kweila/kwela/kwai (r.f. .53). It is also possible
Classificatory Particles in Kilivila
272
to use the general CP kwe (r.f. .27), and, although rather rarely observed, to use the general CP ke (r.f. .03). This semantic domain also offers the expert wicker -worker the CP kauya to refer to the concept 'fish trap, creel'. However, this CP was not elicited in the test, and thus we have no comparative value to describe the position of this CP type within Domain 11. Table 3.4.11b shows that the CPs peta/ta, kweila/kwela/kwai, kwe, and ke3 play the most important roles within Domain 11. Tables 3.4.12a and b attempt to present the dynamics of Domain 12 and provide an evaluation of the CP types that constitute this domain. Table 3.4.12a documents that within Domain 12 there was 1 case of in-domain substitution, affecting the general CP kwe of Domain 4, and 2 cases of intradomain substitution. tWe also observe two cases of off-domain substitution, affecting Domains 3b (quantities) and 5 (measures). Table 3.4.12b presents the CP types that constitute Domain 12 and gives the r.f.s for the production of the individual CP types within the entire domain.
Table 3.4.12a. Domain 12: Yam In-domain substitution
Off-domain substitution tarn
8 .04 28 .16
34 .19
(Domain 4)
kwe
117 +++ .65
0
uva (Domain 5) gula (Domain 3b)
2 .03
1
.01 tetu
Table 3.4.12b. Domain 12: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and infradomain)
1.
2. 3. 4.
kwe 'thing' 117 34 tarn 'sprouting' 0 'basketful of yams' 2 1 telu 'yams'
154
r.f. r.f. r.f. r.f.
.76 .22
.01 .01
1.00
The System of Classificatory Particles in Kilivila
273
On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 12: If speakers want to refer to the concept 'yams, a basketful of yams', they usually use the general CP kwe (r.f. .65). The CP tarn is also produced (r.f. .19; here the speakers have obviously broadened the concept of this CP from 'sprouting yams' to the general concept 'yams'). The use of the special and appropriate zero classifier (r.f. .03) to refer to the concept 'basketful of yams' is almost as rarely observed as the production of the CP tetu (r.f. .01) to refer to the concept 'yams'. However, it must be noted here that the zero classifier is only used with numerals when basketfuls of yams are counted. Table 3.4.12b shows that Domain 12 is dominated by the general CP kwe and that the only other CP type that plays an important role in this domain is tam, with a frame of reference that seems to be broadened to also encompass the general concept 'yams'. I should note here that it may well be that the zero classifier is so closely associated with the ritual dadodiga (the festive filling of the yam houses just before the ceremonious harvest festival period milamald) that this CP is not produced in more profane situations and thus cannot be elicited in a language production test (for a discussion of this CP type, see also the description of Domain 5 given previously). Tables 3.4.13a and b attempt to present the dynamics of Domain 13 and provide an evaluation of the CP types that constitute this domain. Table 3.4.13a documents that within Domain 13 there were 7 cases of in-domain substitution, affecting CPs in Domains 3b (quantities), 4 (general CPs), and 10 (trees, wood, wooden things). Table 3.4.13b presents the CP types that constitute Domain 13 and gives the r.f.s for the production of the individual CP types within the entire domain.
Table 3.4.13a. Domain 13: Part of a foodhouse/a canoe/a creel In-domain substitution (Domain 10) (Domain 4) (Domain 3b)
kel kwe pila
(Domain 3b)
utu
(Domain 10)
bwa
(Domain 4)
kwe
(Domain 10)
kel
53 .29 10
+++
kabisi
78
.43
+++
.06 1
+++
.01 4 .02 4 .02 55 .31 32 .18
+++ +++ +++ +++
liku
71 .39
274
Classificatory Particles in Kilivila
Table 3.4.13b. Domain 13: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2. 3. 4. 5. 6. 7.
8.
kel 'wood' kabisi 'compartment' liku 'compartment' kwe 'thing' bwa 'tree, wood' utu 'part' pila 'part'
lipu 'compartment/creel'
53 + 32 78 71 10 + 55 4 4 1
85 78 71 65 4 4 1 308
r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.28 .25 .23 .21 .01 .01
.003 .993
-/- (= not elicited in the test)
On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 13: If speakers want to refer to the concept 'compartment of a foodhouse, section/division in a foodhouse', they usually use the special and appropriate CP type kabisi (r.f. .43). The CP kel (r.f. .29) is also used in this context, but less frequently. Moreover, it is also possible, although rather rarely observed, to refer to this concept with the CPs kwe (r.f. .06), bwa (r.f. .02), utu (r.f. .02), and pila (r.f. .02). If speakers want to refer to the concepts 'compartment of a canoe, compartment of a foodhouse', they use the special and appropriate CP liku only if they are persons of status (r.f. .39). Otherwise, the general CP kwe (r.f. .31) or, although less frequently, the CP kel (r.f. .18) is used. Domain 13 also offers the expert wicker worker the CP lipu to refer to the concept 'compartment of a creel'. However, this CP was not elicited in the test, and thus we have no comparative value to describe the position of lipu within Domain 13. Table 3.4.13b shows that the CPs kel, kabisi, liku, and kwe play the most important roles within Domain 13. (Tables 3.4.14a and b attempt to present the dynamics of Domain 14 and provide an evaluation of the CP types that constitute this domain. Table 3.4.14a documents that within Domain 14 there was 1 case of intradomain substitution and 4 cases of in-domain substitution, affecting CPs in Domains 4 (general CPs), 8 (qualities), and 16 (road, journey). Table 3.4.14b presents the CPs that constitute Domain 14 and gives the r.f.s for the production of the individual types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 14: If speakers want to refer to the concept 'door entrance', they use the special and appropriate CP type duya/kaduyo only if they are persons of high status (r.f. .05). Only women of high status use the CP va/vaya/vayo (in its first meaning 'door') to refer to this concept (r.f. .05). Speakers usually use the CP kada/keda (r.f. .51), obviously broadening the frame of reference of this
The System of Classificatory Particles in Kilivila
275
Table 3.4.14a. Domain 14: Door, entrance, window In-domain substitution (Domain 16)
kada
(Domain 8)
ya
(Domain 4)
ke
(Domain 4)
kwe
183 .51 45
.12 9 .02 25 .07
+++
duya/kaduyo
19 .05
+++ +++ +++ 17 .05
'a/vaya/vayo
Table 3.4.14b. Domain 14: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2. 3. 4. 5. 6.
kada 'road, track' ya 'flexible, thin' kwe 'thing' duya/kaduyo 'door, entrance' va/vaya/vayo 'window, door' ke3 'inanimates'
183 45 25
19
17 9 298
r.f. r.f. r.f. r.f. r.f. r.f.
.61 .15 .08 .06 .06 .03 .99
CP to encompass the general concept 'passage, way, opening'. It is also possible, although not very often observed for speakers to use the CPs ya (r.f. .12; doors are generally made of palm branches; they are thin and flexible), kwe
(r.f. .07), and he (r.f. .07).
It should be noted here that one of the results of the elicitation test was that the formative duya and kaduyo/kaduya should be regarded as variants of a single CP type. However, I elicited the formatives duya and kaduya/kaduyo separately. Thus, the r.f.s given in the variable rule above are computed on the basis of a possible 360 tokens, in the ideal case. Table 3.4.14b shows that Domain 14 is dominated by the CP kada/keda', however, probably only the general CP ke does not play a role within this domain. Tables 3.4.15a and b attempt to present the dynamics of Domain 15 and provide an evaluation of the CP types that constitute this domain. Table 3.4.15a documents that within Domain 15 there was 1 case of intradomain substitution and 2 cases of in-domain substitution, affecting the general CP kwe of Domain 4.
Classificatory Particles in Kilivila
276
Table 3.4.15a Domain 15: Fire, oven Iti-domain substitution (Domain 4)
kwe
1 +++ .01
kova
158 .88
1 .01
ke4 (Domain 4)
kwe
1 +++ .01
kumla
162 .90
Table 3.4.15b. Domain 15: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
2. 3. 4.
kumla 'earth-oven' kova 'fire' kwe 'thing' ke4 'fire'
162 158 1+ 1 1
162 158 2 1
323
r.f. r.f. r.f. r.f.
.50 .49 .01 .003 1.003
Table 3.4.15b presents the CPs that constitute Domain 15 and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 15: If speakers want to refer to the concept 'fire, fireplace', they generally use the special and appropriate CP kova (r.f. .88). It is also possible, although rather rarely observed, to use the CP ke4, in its meaning 'fire' (r.f. .01) and to use the general CP kwe (r.f. .01). If speakers want to refer to the concept 'earth oven', they generally use the special and appropriate CP kumla (r.f. .90). It is also possible, although rather rarely observed, to use the general CP kwe (r.f. .01) in this context. Table 3.4.15b shows that this domain is dominated by the CPs kumla and kova. Tables 3.4.16a and b attempt to present the dynamics of Domain 16 and provide an evaluation of the CP types that constitute this domain. Table 3.4.16a documents that within Domain 16 there was 1 case of intradomain substitution and 3 cases of in-domain substitution, affecting CPs in Domains 4 (general CPs) and 8 (qualities). There was also 1 case of off-domain substitution, affecting Domain 14 (door, entrance, window); see the previous discussion of Domain 14.
The System of Classificatory Particles in Kilivila
277
Table 3.4.16a. Domain 16: Road, journey In-domain substitution
Off-domain substitution kada/keda
183 .51
152 .84
duya/kaduyo (Domain 14)
.01
(Domain 4)
kwe
(Domain 4)
ke
(Domain 8)
ya
131 .73 1 .01
+++
16
+++
.09
lilo/lola
7 .04
+++
bililo
<4>
kali
<1>
Table 3.4.16b. Domain 16: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1.
2 3. 4. 5.
kada 'road, track' kwe 'thing' ya 'flexible, thin' lilo 'walk, journey' ke 'inanimates'
131 16 7 1
6. 7.
kali 'paddle strike' bililo 'trip'
308 -/- (= not elicited in the test) -/-
152 + 1
153 131 16 7 1
r.f. r.f. r.f. r.f. r.f.
.50 .43 .05 .02 .003 1.003
Table 3.4.16b presents the CPs that constitute Domain 16 and gives the r.f.s for the production of the individual CP type within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 16: If speakers want to refer to the concept 'road, track', they produce the special and appropriate CP kada/keda (r.f. .84). If speakers want to refer to the concept 'walk, journey', they only very rarely use the special and appropriate CP lilo/lola (r.f. .04). Instead, speakers usually use the general CP kwe (r.f. . 73). It is also possible, although not very often observed, for speakers to use the CPs ya (r.f. .09; here the speaker may intend to emphasize the narrowness of most paths on the islands), kada/keda (r.f. .01; here the speaker may intend to emphasize the road one uses on the journey or the trace one follows in one's walk), and the general CP ke (r.f. .01). However, ya and kada/keda in this context seem to be at least somewhat idiosyncratic.
Classificatory Particles in Kilivila
278
Domain 16 also includes the CP bililo to refer to the concept 'trip' and the CP kali to refer to the concept 'paddle strike'. However, these two CPs were not elicited in the test, and thus we have no comparative value to describe their position within Domain 16. Table 3.4.16b shows that only the CPs kada/keda and kwe play a role within Domain 16. Tables 3.4.17a and b attempt to present the dynamics of Domain 17 and provide an evaluation of the CP types that constitute this domain. Table 3.4.17a documents that within Domain 17 there were 2 cases of intradomain substitution and 9 cases of in-domain substitution, affecting CPs in Domains 3b (quantities), 4 (general CPs), and 9 (shape). Table 3.4.17b presents the CP types that constitute Domain 17 and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 17: Table 3.4.17a. Domain 17: Text In-domain substitution megwa 12 .07
(Domain 3b)
pila
(Domain 3b)
utu
67 .37 3
+++
nina
+++
3 .02
.02 (Domain 4)
kwe
61
+++
.34 (Domain 9)
kasa
3 .02
+++
vosi 18 .10
(Domain 9)
kasa
(Domain 4)
ke
(Domain 4)
kwe
(Domain 3b)
pila
(Domain 3b)
utu
10 .06 12 .07 9 .05 91 .51 4 .02
+++
meila
+++ +++ +++
+++
(sec also sisi)
3 .02
The System of Classificatory Particles in Kilivila
279
Table 3.4.17b. Domain 17: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2. 3. 4. 5. 6. 7. 8. 9.
pila 'part, piece' lave 'thing' vosi 'song' kasa 'row, line' megwa 'magic' ke 'inanimates' utu 'part, fragment' nina 'part of a song' meila 'part of a song'
67 + 91 61 + 9 18 3 + 10 12 12 3+4 3 3
158 70 18 13 12 12 7 3 3 296
r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f. r.f.
.53 .24 .06 .04 .04
.04 .02
.01 .01 .99
If speakers want to refer to the general concept 'part(s) of a song', they only very rarely use the special and appropriate CP type nina (r.f. .02). Speakers usually use either the CP pila (r.f. .37) or the general CP kwe (r.f. .34) in this context. It is also possible, although not very often observed for speakers to use the CPs megwa (r.f. .07; here it should be noted that many magical formulae are recited in a way that resembles a song), utu (r.f. .02), and kasa (r.f. .02; here the speaker may intend to refer to a 'row/line of a song'). If speakers want to refer to the concept 'part of a song, of a magic formula', they only very rarely use the special and appropriate CP meila (r.f. .02). Speakers usually use the CP pila (r.f. .51) in this context. It is also possible, although not often observed, for speakers to use the CP vosi, if the speaker is a woman of high status (r.f. .10), to use the general CPs ke (r.f. .07) and kwe (r.f. .05), or to use the CPs kasa (r.f. .06; here the speaker may intend to refer to a 'row/line of a song/magical formula'), and utu (r.f. .02). It should be noted here that Domain 17 also includes the CP sisi to refer to the concept 'division of a magical formula'. However, this meaning was not elicited in the test; I only discovered it during lexicographic work with consultants. Thus, we have no comparative value to describe the position of sisi within Domain 17. Table 3.4.17b shows that Domain 17 is dominated by the CP pila and that the only other CP that plays an important role is the general CP kwe. Tables 3.4.18a and b attempt to present the dynamics of Domain 18 and provide an evaluation of the CP types that constitute this domain. Table 3.4.18a documents that within Domain 18 there was 1 case of intradomain substitution and 1 case of in-domain substitution, affecting the general CP kwe of Domain 4. Table 3.4.18b presents the CP types that constitute this domain and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 18:
280
Classificatory Particles in Kilivila
Table 3.4.18a. Domain 18: Ritual items In-domain substitution beku 2 .01
(Domain 4)
kwe
2 +++ .01
kai
136 .76
sam
<7>
Table 3.4.18b. Domain 18: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain)
1. 2. 3.
kai 'stone blade' beku 'stone blade' kwe 'thing'
136 2 2 140
4.
sam 'ginger'
r.f. r.f. r.f.
.97 .01 .01 .99
-/- (= not elicited in the test)
If speakers want to refer to the concept 'stone blade', they generally produce the special and appropriate CP kai (r.f. .76). It is also possible, although rather rarely observed, to use the synonymous CP beku (r.f. .01) or the general CP kwe (r.f. .01). This domain also includes the CP sam to refer to the concept 'ginger' in play accompanying verses; ginger plays an important role in Trobriand magic (see section 3.3.4.1). However, this CP type was not elicited in the text, and thus we have no comparative value to describe the position of sam within Domain 18. Table 3.4.18b shows that Domain 18 is dominated by the CP kai. Tables 3.4.19a and b attempt to present the dynamics of Domain 19 and provide an evaluation of the CP types that constitute this domain. Table 3.4.19a documents that within Domain 19 there was 1 case of intradomain substitution and 1 case of in-domain substitution, affecting the general CP kwe of Domain 4. There was also 1 case of off-domain substitution, affecting Domain 9 (shape). Table 3.4.19b presents the CP types that constitute Domain 19 and gives the r.f.s for the production of the individual CP types within the entire domain.
The System of Classificatory Particles in Kilivila
281
Table 3.4.19a. Domain 19: Dress, adornment In-domain substitution
Off-domain substitution doba 1 .01
(Domain 4)
kwe
145 +++ .81
kudul vakala
6 .03
2 .01
gili (Domain 9)
Table 3.4.19b. Domain 19: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain) 1. 2. 3.
kwe 'thing' kudul 'band of fibers' doba 'grass-skirt'
145 6 1
4.
vakala 'belt'
0 (= not realized within this domain; see Domain 9 above)
152
r.f. r.f. r.f.
.95 .04 .01 1.00
On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 19: If speakers want to refer to the concept 'band of fibers' (at the waistband of a grass skirt), they usually use the general CP kwe (r.f. .81). The use of the special and appropriate CP kudul (r.f. .03) is almost as rare as the more general CP doba (r.f. .01) in this context. This domain also includes the CP vakala to refer to the concept 'belt of spondylus shell discs'. However, this CP was only used in a rather idiosyncratic way to refer to the concept 'row' in Domain 9. Table 3.4.19b shows that the general CP kwe dominates Domain 19. Tables 3.4.20a and b attempt to present the dynamics of Domain 20 and provide an evaluation of the CP types that constitute this domain. Table 3.4.20a documents that within Domain 20 there was 1 case of in-domain substitution, affecting the general CP kwe of Domain 4. Table 3.4.20b presents the CP types that constitute Domain 20 and gives the r.f.s for the production of the individual CP types within the entire domain. On the basis of these two tables, I have formulated the following variable rules that speakers use in their production of CPs in Domain 20: If speakers want to refer to the concept 'name', they generally produce the
282
Classificatory Particles in Kilivila
Table 3.4.20a Domain 20: Names In-domain substitution (Domain 4)
kwe
6 +++ .03
iga/yegila (142) (7)
149 .83
(.79) (.04)
Table 3.4.20b. Domain 20: Relative frequencies interpreted as figures for the evaluation of CP types (in-domain and intradomain) 1. 2.
iga/yegila 'name' kwe 'thing'
142 + 7 6
149 6 155
r.f. r.f.
.96 .04 1.00
special and appropriate CP iga/yegila (r.f. .83). It is also possible although rather rarely observed, to use the general CP kwe (r.f. .03) in this context. Table 3.4.20b shows that Domain 20 is dominated by the CP iga/yegila. Now that I have presented the dynamics of all 20 semantic domains and the rules that speakers may use in their production of CPs, it is time to reflect once again on the proceeding proposed in this section and to further define the concept "dynamic semantic domain". My goal in this section has been to present the dynamics of the 20 semantic domains and to formulate language production rules that predict a speaker's choice of a certain CP type to refer to certain nominally expressed concept(s) of referents in the extralinguistic world. To reach this goal, I first defined the semantic domains by grouping together certain CPs, based primarily on common sense considerations that took into account ethnographic and sociological information about the speech community. This procedure provided the basis for describing what actually happens if a certain CP type is produced to refer to a given semantic concept. The observation of the processes that I have called intradomain and in-domain substitution and the weighting and evaluation of these processes by computing the relative frequency with which a certain CP is used within the domain as a whole resulted, on the one hand, in the formulation of variable rules, which I interpreted as rules that speakers use in their production of a certain CP type of the semantic domain, and, on the other hand, in an insight into the evaluated distribution of the individual CP types within each semantic domain.
The System of Classificatory Particles in Kilivila
283
The logic inherent in this proceeding can be summarized as follows: If speakers want to transform a given semantic concept into an appropriate CP to refer to this concept, they first must assign the semantic concept to the semantic domain that encompasses this concept. Next, speakers must call in and activate the variable rule or rules that they use in their production of a certain CP type of this domain that refers to the concept given. If the rule allows the production of more than one CP type in this context, they must decide which of the possible CP types is most appropriate for their purposes pursued in referring to the given concept. If they are persons of high status and if the variable rule activated includes a CP type that functions as a sociolinguistic variable, they must decide whether they want to use this CP with its sociolinguistic implications. All these decision processes then lead to the production of the CP that the speakers assume to be most appropriate to refer to the given semantic concept and for the means and ends they want to pursue with their verbal reference. Thus, my attempt to predict a speakers's choice of a certain CP to refer to a certain nominal concept (or concepts) also emphasizes that, among other things, CPs must be understood as formatives that can be used strategically to serve certain means and ends a speaker wants to pursue and express. Moreover, my proceeding also emphasizes that semantic domains as those constituted by the CPs are not static at all. These semantic domains are dynamic and interact with each other. They can be understood as "program clusters, procedures, scripts" or "functional pathways" (see Pribram 1987:7-12) that speakers employ and rely on in their speech production.11 Before I discuss this important point in more detail (see chapter 6), I will summarize the data from Tables 3.4.1.1b to 3.4.20b (see Table 3.4.21). Table 3.4.21 shows that the 20 semantic domains are connected or "short-circuited" by 18 CP types (two of which are represented with two different meanings each). As might be expected, the general CP kwe (= kwe1) plays a role in 19 different semantic domains. The only domain where kwe does not play a role is Domain 2; however, here kwe in its meaning 'shells and clams' (= kwe2) is one of the domain-constituting CPs. The other general CP ke ( = ke3) plays a role in 11 different semantic domains. The CP ya connects 8 and the CPs pila and utu each connect 7 semantic domains. The CP bwa plays a role in 5 domains, and the CP kasa occurs in 4 domains. The CPs ke1, tam, and vili connect 3 domains each. Finally, the CPs bubwa, kada, meila, na ( = na1+na2), oyla, si, sisi, to/te ( = to/te2), and vakala play a role in 2 domains each. 11 In a heuristic phase of the analysis of the dynamics of these domains, I noted the cases of in-domain, intra-domain, and off-domain substitution for all 20 domains on a rather large sheet of paper. The result was a kind of "drawing" that looked like a mycelium. Any reader whose enthusiasm about Kilivila CPs equals that of the author can easily redraw this picture by putting the semantic domains described together and then connecting them according to the notation conventions marking the cases of in-domain and off-domain substitution.
284
Classificatory Particles in Kilivila
Table 3.4.21. CPs and their semantic domains Domains
CP beku bililo bogi bubu/bobu bubwa
18 16
6 8
3b 8
buda
3a 3b
bukwa bwa
3b
bwalita deli doba duli dumia duya/kaduyo gili giwi guba gudi gula gum iga/yegila kabisi kabulo kabulo1 kabulo2 kada/keda kai ka'i kala kali kapwa kaaa
3b 8 9 10
13 7 3a 19 3b 7 14 9 8 3b la
3b 3b 20 13
7 7 14 16 18 1b 6 16 3b 3a 3b 8
Ritual items Road, journey Time Qualities Quantities Qualities Quantities Quantities Quantities Quantities Qualities Shape Trees, wood, wooden things Part of a foodhouse/canoe/creel Place Quantities Dress, adornment Quantities Place Entrance, window Shape Qualities Quantities Persons Quantities Quantities Names Part of a foodhouse/canoe/ creel Place Place Entrance, window Road, journey Ritual items Body parts Time Road, journey Quantities Quantities Quantities Qualities
No. of domains
l 1 1 1
r.f. .01 -/.02 .08
2
.004 .02
1
.22 .001
1
5 1 1 1 1
.001 .01 .02 .005 .01 .01 .01 .14 .01 .05
1
-1.06 .02
1
.12
1
.001
1
.29 .03
1 1
1 1 1 1
1 1 2 1 1 1 1 1
.04 .96 .25
.003 .02 .61 .50 .97 .27 .01
-/.02 .001 .001 .003
The System of Classificatory Particles in Kilivila
285
Table 3.4.21. (continued) CP
Domains 9
kauya kavi ke ke1 ke2 ke3
17 11 11
3b 10 13 9
1b
3a 3b 4 5 7
8 9 11 14 16 17 ke4
kila kova koya/kwoya kubila/kwabila kudu kudu1 kudu2 kumla kwe kwe1
15 3b 15
7 7
19 1b 15
1b 3a 3b
4
5 6 7 8 9 10 11 12 13 14 15 16
Shape Texts Utensils Utensils Quantities Trees, wood, wooden things Part of a foodhouse/canoe/creel Shape Body parts Quantities Quantities General CPs Measures Place Qualities Shape Utensils Entrance, window Road, journey Texts Fire, oven Quantities Fire, oven Place Place Dress, adornment Body parts Fire, oven Body parts Quantities Quantities General CPs Measures Time Place Qualities Shape Trees, wood, wooden things Utensils Yams Part of a foodhouse/canoe/creel Entrance, window Fire, oven Road, journey
No. Of domains 4 1 1
3 1
11 1 1 1 1 1 1 1 1
r.f. .24 .04 -/.01 .13 .82 .28 .18 .02 .001 .15 -/-; 0 .70 .10 .26 .003 .12 .03 .003 .04 .003
.003 .49 .03 .10 .04 .06 .50 .15 .26 .24 1.00 .05 .91 .70 .07 .12 .002 .12 .76 .21 .08 .01 .43
286
Classificatory Particles in Kilivila
Table 3.4.21. (continued) CP
kwe2 kweya kwoila/kweya liku lila lilo/lola lipu luba luva luva1 Iuva2 megwa meila mmwa na na1 na2 na3 na4 na5 na6 nigwa nina notu/nutu nunu oyla pila
Domains 17 18 19 20 2
1b 11 13 10 16 13
3b 10 3b 17 6 17 3b la 3a 2 3a 7 10 1 1
8 17 9 7 3a 11 3b 6 7
8 11 13 po'ula pwanina/pona sa
17 3b 8 3b
Texts Ritual items Dress, adornment Names Animals Body parts Utensils Part of a foodhouse/canoe/creel Trees, wood Road, journey Part of a foodhouse/canoe/creel Quantities Trees, wood, wooden things Quantities Texts Time Texts Quantities Persons Quantities Animals Quantities Place Wooden things Persons Persons Qualities Texts Shape Place Quantities Utensils Quantities Time Place Qualities Utensils Part of a foodhouse/canoe/creel Texts Quantities Qualities Quantities
No. of domains
19 1 1 1 1 1 1 1 1 1 1 1 2 1 2 2 1 1 1 1 1 1 1 1 2
7 1 1 1
r.f. .24 .01 .95 .04 -/.47 .25 .23
0 .02 .-/.04 .004 .004 .04 -/.01 .01 .35 .001 1.00 .17 -/-/-/-/.001 .01 .18 .02 0 0 .04 .002 .001 .004 .01 .003 .53 .03 .16 .01
The System of Classificatory Particles in Kilivila
287
Table 3.4.21. (continued) Domains
CP sam si sipu sisi siva siwa suya ta/peta
tam
tetu to/te to/te1 tolte2
tuta
utu
uva va/vaya/vayo vakala vili
vosi wela
ya
18
Ritual items
No. Of domains
l
r.f.
-/.001 .02
3b 8
Quantities Qualities
3b 8 10 6
Quantities
Time
2 1
7 3a 11 3b 5
Place
1
.04 .12 .16 .01 0
Quantities
1
.13
Utensils
1
Qualities Trees, wood, wooden things
12
Quantities Measures Yams
12
Yams
la la 1b 3a 6
Persons Persons Body parts Quantities
3b 6 7 8 9
Quantities Time Place Qualities Shape Part of a foodhouse/canoe/creel Texts
13 17 5
Time
2 1
3 1 1 2 1
14
Entrance, window
7 1 1
9 19 3b 8 9
Shape Dress, adornment
2
Quantities Qualities Shape
17
Texts
3 1
Quantities
1
Body parts Quantities Qualities Shape Trees, wood, wooden things Utensils Entrance, window Road, journey
8
3a 1b 3b 8 9 10 11 14 16
Measures
.42 .01 .05 .22 .01 -/.36 .01 .02 .01 .10 .01 .01 .04 .06 .01 .02 .20
0
.06 .003
.002 .08 .005 .06 .06 .002 .04 .02 .17 .01 .003 .15 .05
Classificatory Particles in Kilivila
288
Table 3.4.21. (continued) Domains
CP yam1
yam2/yuma1/yama yeni yulai/yule yuma2 yuva 0
6 1b 3b 3b 5 3a 12
Time Body parts Quantities Quantities Measures Quantities Yams
No. Of domains
r.f.
1
.03
1
.03
1 1 1
-/-
.004 -/-
1
.01
1
.01
-/- = not elicited in the test; 0 = not produced; r.f. = relative frequency.
Thus, we can record that roughly a fifth12 of the 88 CPs that make up the 20 semantic domains are decisive for the domain-connecting dynamics of this linguistic phenomenon. Now that I have highlighted the CP types that are most important for the dynamics of the CP system, I also want to answer the question of whether there is a hierarchical order for the 20 semantic domains. So far, I have presented the semantic domains in a random order in which all domains were considered to be of equal rank. However, if we again examine the semantic domains with respect to the question of a possible hierarchical order, we can hardly refrain from hypothesizing such a hierarchy. The problem with this hypothesis, however, is that we must devise a way to express the hierarchical relations between the semantic domains adequately. To solve this problem, we must set out once more on the rather uncertain path of speculation, as follows: So far, I have been using absolute and relative frequencies to describe subsets of the Kilivila CPs. Continuing with this methodological device, I assume that the sum of all the produced tokens of the CPs that constitute a semantic domain can be taken as an indication of the importance of that domain. However, the number of CPs that constitute the 20 semantic domains varies from 1 to 34. Thus, to take this variable into consideration, I have divided the sum of the number of CP tokens produced within a domain by the number of CP types produced within that domain. The result is an index that I regard as the measure that expresses the hierarchical status of the domain within the Kilivila speech community. This admittedly rather simple and unsophisticated proceeding leads to the results presented in Table 3.4.22. This table shows the 20 semantic domains in the hierarchical order expressed by the "hierarchy index". 12 We can ignore the CPs meila, oyla, and vakala here. Although they connect two domains, as Table 3.4.21 shows, they are not realized in one or both of these domains. Thus, we are left with 15 CP types (17% of the 88 CP types described) that are decisive for the domain-connecting dynamics.
The System of Classificatory Particles in Kilivila
289
Table 3.4.22. The 20 semantic domains of Kilivila CPs in hierarchical order
I
II
III
1.
Persons General CPs Animals Trees, wooden things
5. 6. 7.
Persons, body parts Place Quantities Quantities (-animate) Fire, oven
2. 3. 4.
8. 9.
Hierarchy index
Domain no.
Domain group
la 4 2 10 1a and b 7 3a and b
15
495: 3 164: 1 162: 1 903: 6 950: 910: 2,790: 2,109: 323:
10 10 34
2 8
20 6
14.
Names Time Quantities (+animate) Road, journey Qualities
8
155: 609: 681: 308: 898:
15. 16. 17. 18. 19. 20. 21.
Shape Body parts Utensils Dress, adornment Door, entrance, window Ritual items Part of a foodhouse/a canoe/a creel
9 1b 11 19 14 18 13
585: 455: 377: 152: 298: 140: 308:
22. 23. 24.
Measures Yams Texts
5 12 17
10. 11. 12. 13
3a 16
26 4
165.0 164.0 162.0 150.5 95.0 91.0 82.1 81.1 80.8
15
77.5 76.1 61.9 61.6 59.9
10 8 7 3 6 3 7
58.5 56.9 53.9 50.7 49.7 46.7 44.0
157: 4 154: 4 296: 9
39.3 38.5 32.9
11 5
IV
V
The computing of the hierarchy index for the 20 semantic domains orders the domains in such a way that they can be roughly classified into five different groups (I-V).13 If we consider subdomain la as a domain, the semantic domain covering the concept "persons" is highest in the hierarchy. This domain is followed by the domains covering the concepts "general CPs", "animals", and "trees, wood, wooden things". These four domains form a group of the most important semantic domains. The second group contains subdomain la and subdomain 1b, encompassing the concepts "persons and body parts", and the domains that cover the concepts "place", "quantities", and "fire, oven". If we consider the subdomain 3b as a domain, encompassing the concept "quantities, with generally inanimate referents", then it also belongs to this second group. 13 These numbers indicating the five groups of semantic domains should not be confused with the Roman numerals used to indicate the five age groups of the consultants (see section 3.2).
290
Classificatory Particles in Kilivila
A third group, which may be considered as occupying an intermediate position with respect to rank and importance of the semantic domains, includes the domains that cover the concepts "names", "time", "road, journey", and "qualities". Domain 3a, covering the concept "quantities with generally animate referents", also belongs to this third group of semantic domains. The group of less important semantic domains (group IV) consists of the domains covering the concepts "shape", "utensils", "dress, adornment", "door, entrance, window", "ritual items", and "part of a foodhouse/a canoe/a creel". Considering subdomain 1b as a domain, encompassing the concept "body parts", it also belongs to this fourth group. The fifth and last group of semantic domains is formed by the least important domains, covering the concepts "measures", "yams", and "texts". More detailed comments on Table 3.4.22 will be given in chapter 5. However, I will state one caveat here and briefly remark on one of the semantic domains with respect to its status within this computed hierarchy. First, the caveat: I must emphasize again that the hierarchy presented in Table 3.4.22 is based on speculation. Although I include this table in the arguments given in the next section and in chapter 5, this warning should be kept in mind. The second remark I want to make here refers to Semantic domain 12, which covers the concept "yams". Those familiar with Trobriand ethnography may doubt that this domain holds the second lowest rank within the hierarchy, since yams are one of the most important factors in Trobriand society (see Malinowski 1935; Weiner 1976, 1988; Senft 1985d:476, 490, 1987d). However, in my previous discussion of Domain 12, I noted that at least the zero classifier may be so closely linked with ritual and ceremony that it is not produced in profane situations and, thus, cannot be elicited in a language production test. This may explain the rather low rank of Domain 12; it is probably an artifact of the computation of the hierarchy index. I want to again state that we are on safe groud if we assume that all semantic domains are of equal rank. However, if we dare to speculate that there is a hierarchy of semantic domains, we may be rewarded by some further ideas and maybe even insights. Before I proceed with an argument along these lines, one question remains with respect to the acquisition of the CPs. This question will be discussed in the next section.
3.3.4.3 A Flashback: Why Are the Individual CPs Acquired in a Certain Order? I have tried to describe the processes of language change affecting the CPs and the dynamics of the semantic domains constituted by the CPs, and I have speculated on a possible hierarchical order of these domains. I can now attempt to answer the question that was raised in section 3.3.2.4 but could not be answered at that time because of lack of information. As Chin has stated, "to adequately understand acquisition studies on numeral classifiers,
The System of Classificatory Particles in Kilivila
291
one needs to understand the adult usage of the system" (Chin 1989:87; see also p. 105). The arguments in this section refer back to Table 3.3.5 in section 3.3.2.4 and to the first interpretation of this table given there. Moreover, they rely on the results presented in sections 3.3.3 and 3.3.4. Table 3.3.5 (see pp. 191-192) shows that CPs that distinguish gender in human beings (to/te2, na1) are acquired first. They come first because their frame of reference is distinct and because they constitute the highest ranking semantic domain of the domain hierarchy. To be able to refer to persons is obviously so important that the CPs needed for this are acquired first (I/1, Table 3.3.5). General CPs (kwe, ke3) are acquired next because the semantic domain of these CPs occupies the second highest rank within the domain hierarchy. General CPs allow a grammatically correct, although semantically void, reference to everything, and soon probably also to everyone (I/2, Table 3.3.5; see also Adams 1989:177). The CPs that are acquired next (kova, na2, ta, ya, kumla, pwanina/pona, gudi, and sisi) have, with the exception of ya, a distinct frame of reference. They all denote concrete, specific objects or features of objects; moreover, most of them are in semantic domains that are the most important within the domain hierarchy. The exceptions are ta, ya, and pwanina/pona; they occur in semantic domains that are intermediate in the domain hierarchy. However, ta and pwanina/pona refer unequivocally to concrete objects, and ya plays a rather important role with respect to the dynamics of the semantic domains (see Table 3.4.21). Therefore, it seems necessary for establishing the whole system of semantic domains that this CP type is acquired at a rather early stage of the acquisition process (I/3, Table 3.3.5). The same explanation holds for the CPs that are acquired in the fourth phase of the acquisition process (I/4, Table 3.3.5) by girls and boys, regardless of whether some of them are acquired earlier by boys than by girls. The CPs utu, bwa, kada, si, and kasa play a rather important role with respect to the dynamics of the semantic domains (see Table 3.4.21). This function seems to be so important that even a CP like si, which is rarely produced in actual speech, is nevertheless acquired rather early. The CP iga has a distinct frame of reference and occurs in one of the domains that is intermediate in the domain hierarchy. As for the CPs that are acquired by boys in this stage of the acquisition process, we note once again that pila and bubwa play a rather important role with respect to the dynamics of the semantic domains (see Table 3.4.21). The CPs duli, kai, kweya, notu/nutu, and kweila/kwela all have a distinct frame of reference. The CPs buda, dumia, bukwa, kapwa, luba, luva, and sa all occur in semantic domains that are important within the domain hierarchy. Obviously at this stage of the acquisition process, it becomes necessary for boys to be able to refer to concepts that denote quantity. What is striking, however, is that among these CPs we find two that are almost obsolete (dumia, bukwa), two that play a secondary role in actual speech (luba, luva), and three that function as sociolinguistic variables (bubwa, kapwa, sa). This observation of the acquisition of almost obsolete
292
Classificatory Particles in Kilivila
CPs and CPs that play a secondary role in actual speech production may be a coincidental result. However, the fact that the acquisition of CPs that function as sociolinguistic variables is noted at this stage of the acquisition process emphasizes the important societal role of this system of formatives. The last observation also holds for the CPs that are acquired only by girls at this stage of the acquisition process. The CP lilo/lola is almost obsolete, and the CPs kaduyo, kwoya/koya, va, and yuma function as sociolinguistic variables; moreover, the CPs va and yuma are found only in the speech of women of status. For a possible explanation of the gender-specific differences in the CP acquistion process, the reader is referred to section 3.3.2.3. In the fifth phase of the acquisition process (II/5, Table 3.3.5), we note the acquistion of the following CPs by boys and girls (regardless of whether some are acquired earlier by boys than by girls): giwi, sipu, kubila, and deli all have a distinct frame of reference; bubu, kabisi, and gula are also acquired at this stage of the acquisition process, although they play a secondary role in actual speech. What is most striking here is that the CP kabisi, which occurs in a semantic domain that is less important, is acquired at this stage. This seems to be coincidental. The CPs liku and bogi are also acquired in this fifth phase. They function as sociolinguistic variables, a result that again emphasizes the societal role of Kilivila CPs. The CP gum, which is also acquired at this stage, ocurs in a semantic domain that is more important within the domain hierarchy. With the exception of the CP kabisi, all these CPs are in domains that are important or intermediate within the domain hierarchy. As for the CPs that are acquired only by boys at this stage, the following can be noted: tam, po'ula, and vili play a rather secondary role in actual speech; however, the frames of reference of po'ula and vili are distinct. Moreover, tam and vili play a rather important role with respect to the dynamics of the semantic domains (see Table 3.4.21). The CPs yam, kabulo2, and mmwa function as sociolinguistic variables, again an indication for the importance of these formatives within the highly stratified Trobriand society. That the CP megwa, belonging to a rather technical language variety and occurring in the domain that ranks lowest in the domain hierarchy, is acquired at this stage may be coincidental; however, the consultant producing this CP (consultant 24) will probably inherit important garden and canoe magic from his relatives; thus, this result may reflect the fact that it is with children belonging to age group II (8-14 years) where future magic-specialists are already beginning their education towards a career. As for the CPs that are acquired only by girls at this stage, the following can be noted: suya has a distinct frame of reference; kila, kudu2, nunu, and wela all function as sociolinguistic variables; kila especially marks women of high status. With the exception of tam and megwa, all CPs that are acquired either only by boys or only by girls at this stage occur in semantic domains that are important or intermediate within the domain hierarchy. For a possible explanation of gender-specific differences in the CP acquisition process, see section 3.3.2.3.
The System of Classiflcatory Particles in Kilivila
293
In the final phase of the acquisition process (III/6, Table 3.3.5), the variants yegila and peta of the CP types iga/yegila and ta/peta are acquired. Their frame of reference is distinct; however, their acquisition at this stage seems to be nothing more than a somewhat "belated echo" of the much earlier acquired variants iga (I/4) and ta (I/3). The CP tuta is rather rarely produced but has a distinct frame of reference. The CPs bwalita and vosi function as sociolinguistic variables. As for the CPs that are acquired only by boys at this stage, the following can be noted. The CP yulai/yule is rather rarely produced, but has a distinct frame of reference. The CPs duya, vaya, uva, and gili function as sociolinguistic variables. Finally, it is at this stage that the zero classifier 0 is acquired by boys. The CP types beku, vakala, and yuva are acquired only by girls at this stage; these CPs are rather rarely produced; thus, this result may be somewhat coincidental. In this phase of the acquisition process, we note that the consultants also start to acquire CPs that occur in semantic domains of lesser importance within the domain hierarchy. As for the CP types that were produced by consultants who were finished their first language acquisition (IV+ V, Table 3.3.5) we find that most of these CPs occur in the less important semantic domains within the domain hierarchy. For example, kauya, lipu, nigwa, and yeni mark technical language varieties; it is plausible to include the CPs kudu1 and doba, as well as nina and meila, in this subset of CPs that mark technical language varieties. To summarize, the following reasons can be given to explain the order in which Kilivila CPs are acquired: To differentiate people according to their gender and to refer to things in a grammatically correct way is obviously so necessary for children in Trobriand society that the CPs na1, to/te2, kwe and ke3 are acquired first. For all other CPs, the following factors play a determinant role: (1) frame of reference, (2) role in the dynamics of the CP system, and (3) rank of the semantic domain containing the CP in the domain hierarchy. On the basis of these results, the following general trends, crucial for the acquisition process, can be formulated: 1. The more distinct the frame of reference for a certain CP, the earlier it is acquired. 2. CPs that play an important role in the dynamics of the CP system as a whole (because they connect two or more semantic domains) are acquired earlier than CPs that are not important with respect to these dynamics. 3. The higher the semantic domain containing a CP within the domain hierarchy, the earlier this CP is acquired. 4. CPs that function as sociolinguistic variables are acquired in the phases that characterize the first language acquisition process. Thus, children belonging to clans of high status will acquire these CPs before or during
294
Classificatory Particles in Kilivila puberty (see also section 3.3.2.1). For these CPs, the first three general trends also hold true.14
It should be emphasized that the first three trends cannot be observed separately; they are intertwined and interact with each other. I will now briefly reflect on some considerations about this topic by Tversky (1986) and Tyler (1969). At the end of section 3.3.2.6, I quoted Tversky (1986:72) and Tyler (1969:19), emphasizing that we can use their propositions as a starting point for our attempt to explain the order in which Kilivila CPs are acquired. Now we can understand Tyler's reservations about the common assumption that a "child first acquires concepts for concrete objects and gradually expands its semantic domain by extension and generalization to include more abstract conceptions" (Tyler 1969:19). This assumption and Tversky's statement that perceptually based groupings precede functionally based groupings in the categorization of children (see Tversky 1986:72) are insights that, considering the results of the Kilivila CP data analysis presented here, cover only one aspect of the intricate and complex acquisition process of categorizing formatives. As this section demonstrates, the order in which the categorizing Kilivila CPs are acquired depends on a number of different, interacting factors and variables.
14 The results that led to the formulation of this general trend, which is crucial for the CP acquisition process, emphasize once again how persons of rank within the highly stratified Trobriand clan and subclan society use some of these formatives to mark their societal status. That children who will occupy a high-ranking social standing learn to handle this CP system may serve as a good indication for the fact that the societal function of some of these CPs is a linguistic tool that persons of status consciously use. Being aware of the function of these CPs, they pass this information on to their successors, making sure that they acquire these CPs in childhood during first language acquisition.
Chapter
4 On the Validity of Some of the Presented Results: Six Years Later—A Restudy The results presented in section 3.3 are based on the data I gathered in 1982 and 1983. I started analyzing these data in 1988 and finished in the spring of 1989. Thus, I dealt with this complex system of CPs without having been immersed in the Kilivila speech community for almost six years. After such a long period of separation from the language and the speech community, it goes without saying, at least for the empirical linguist doing field research, that researchers may be somewhat suspicious about their findings, fearing that they may be artifacts produced in their study, far from the reality of the language and speech community that is the object of study. The only way to overcome this "suspicious feeling" is to do field work again and to test the results of one's analyses with the native speakers of the language studied, in keeping with Aristotle's statement: . . . truth is assessed in the light of the facts of actual life; because it is in these that the decisive factor lies. So we must bring what we have already said to the test of facts of life; and if it accords with the facts, we can accept it, but if it conflicts with them we must regard it as no more than a theory. ARISTOTLE, THE NICOMACHEAN ETHICS, Book X, 9, 1179a, 4-26
Thus, In May 1989 I revisited Papua New Guinea for four months to continue my field research on the Trobriand Islands. I conducted a restudy of the system of CPs and checked with my consultants in Tauwema village on Kaile'una Island the following four aspects of the results of my analyses: 1. Do new observations verify or falsify the results I noted with respect to the general temporal progress of the CP acquisition process? 295
296
Classificatory Particles in Kilivila
2. Do Kilivila native speakers accept the semantic domains constituted by the CP types proposed in section 3.3.4.2 or do they modify these domains either by rejecting certain CP types in the domains (as proposed in my analyses) or by adding other CP types? 3. Do Kilivila native speakers accept or reject or modify the variable rules formulated in section 3.3.4.2? Or, to phrase it differently: Do Kilivila native speakers producing a certain CP type referring to a certain nominal concept actually use the variable rules formulated in section 3.3.4.2? 4. Do Kilivila native speakers accept the weighting and ordering of the semantic domains as expressed by the hierarchy index in Table 3.4.22 in section 3.3.4.2 or do they modify and change this weighting? This chapter presents the methods of data gathering that I used and the results of this restudy. In the followong paragraphs, I first briefly comment on the methods used and the results found with respect to the first and the fourth aspect of my restudy and then present the methods used and the results found with respect to the second and third aspect of this restudy. To gather new data on the general temporal progress of the CP acquisition process, I initially wanted to put a small bag with a tape recorder around the neck of our two- and four-year-old children to record verbal exchanges of Trobriand children with whom they played. Although our children managed to get accepted within the children's group of our village sector 'Vaseda' after three weeks, my plan to gather data failed completely: the children most often played on the reef or at the beach and the rather wild games they played there would have damaged the recorders sooner or later. However, among our equipment we had a swing with a good rope to fasten it. Our neighbors, seeing this rope, erected a wooden supporting structure for the swing just in front of our house immediately after our arrival (and after a few weeks they exchanged, with our consent, of course, the new rope for an old one, using the better material for their canoe on one of their travels to Simsim Island). The swing was frequently used during the following three months; its location allowed me to sit in our house and observe and overhear the children's conversation without disturbing or irritating them. Observations were made at random and lasted between four and thirty-eight minutes. During these sessions, I noted the names of the children and the CP types and tokens they produced. I will not present these data in detail; for my purposes here, it suffices to record that these observations verified, without any exceptions, the results presented in section 3.3.2.1 with respect to the general temporal progress of the CP acquisition process. To answer the question of whether Kilivila native speakers accept, modify, or change the weighting and ordering of the semantic domains as expressed by the hierarchy index in Table 3.4.22, I developed the following test: I approached at random 24 consultants (12 female, 12 male) of four different age groups (see appendix C) and asked them to order a set of 24 drawings
Six Years Later—A Restudy
297
on postcards; these drawings were meant to symbolize the 24 semantic domains I found in my 1983 Kilivila CP data. The instruction given was the following: Kulakisi pikisi makala kudoki bukulivala tutatuta kena mimilisi wala. Avaka bukulivala tutatuta bikugwa, avaka mimilisi wala alavigimkoila. (Free translation: 'Order the pictures in the way you think you always or rather rarely talk [about the concepts presented]. What you always talk about comes first, what [you] rather rarely [talk about] comes last'). I must confess that my efforts to analyze these data were in vain; I could not find any systematic pattern in my consultants' responses to this test. This may be because they perceived the drawings presented as stimuli in very idiosyncratic ways. Moreover, I cannot control whether the consultants reacted to the topics symbolized in the drawings or whether they responded to aesthetic aspects of the drawings or whether some other factors influenced them. Thus, the question raised here must remain unanswered. With this confession of a poor, inadequate, and therefore unsuccessful test design I have to go on to the presentation of the methods pursued, data gathered, and results found with respect to checking the other two aspects of my restudy that concern the semantic domains postulated and the variable rules of CP production formulated. My next task was to ascertain if Kilivila native speakers accept the semantic domains proposed in section 3.3.4.2 or modify these domains by rejecting certain CP types in the domains or by adding other CP types. Therefore, I approached at random 30 consultants (15 female and 15 male) in five different age groups (see appendix C) and asked them whether they would like to cooperate. All 30 responded positively. I then used the following elicitation format: Olopola kwetala biga dabigasi esisusi sitana makala "na" pela "na-manabweta vivila", makala "to" pela "to-manabweta taw". Kidamwa magigu balivala avaka avaka bakebiga pela .... (Free translation: 'Within certain words of our language there are bits like na for na-manabweta vivila ['female-beautiful girl'] or like to for to-manabweta tau ['male-beautiful man']. If I want to speak about something I will say referring to ....') I then finished the sentence with a noun or a noun phrase describing the 20 semantic domains; for example, tommota kumwedona 'all people', vovosi tommota 'people's body (parts)', mauna 'animals'. Next, I presented all the classifiers in that domain, together with an appropriate noun or a noun phrase; for example, Esisu 'gudi' pela gwadi kena gugwadi. (Free translation: There is gudi for 'child' or 'children"). After the presentation of all CPs in the domain, I asked the following questions: 1. Kudoki sena bwenal (Free translation: 'Do you think this is all right?') 2. Kudoki kunanamsa akigagi kena ayogagi kwetala kwetala kena bigatala makala kwetala biga sitana? (Free translation: 'Do you think that I forgot something or made a mistake (somewhere) with some part, with one of these bits of words?')
298
Classificatory Particles in Kilivila
If the second question was answered positively, I continued with the following questions: Avaka akigagi? Avaka ayogagi? (Free translation: 'What did I forget? Where did I make a mistake?') I noted the answer, and I asked again: Kudoki mabigasina sitana kwetala biga lalivala oluvi kumwedona biga Tauwema mokita? Kena asula kidamwa alivala makala? Kidamwa asula, magigu bukulivala ambe asula. (Free translation: 'Do you think that all these parts of our language I mentioned before [represent] correct Tauwema speech? Or do I make a mistake when I speak in this way? If I make a mistake, I want you to tell me where I make it'). All the 30 interviews with my consultants were tape recorded; moreover, I also wrote down the reactions of my consultants. I do not think it is necessary to present the complete elicitation test here; rather, this sample presentation should suffice. My analysis of the data collected using this elicitation test showed the following results: First, the responses of the children 4-7 years old confirmed my earlier results with respect to the CP acquisition process. The children could only respond to the CPs that they had already acquired (see section 3.3.2.4, Table 3.3.5), of course. Thus, the responses of this age group were not useful in answering the questions raised here. With the consultants in age groups II-V, the following interesting trend was noted: the observed overgeneralization of the two general CPs kwe and ke was rejected by a number of consultants. Thus, for Domain 1b (body parts), 3 consultants rejected the CP kwe and 8 consultants rejected the CP ke. For Domain 3a (quantities (+ animate)), 9 consultants rejected the CP kwe and 13 consultants rejected the CP ke. For Domain 7, (place), 7 consultants rejected the CP ke. For Domain 10 (trees, wood, wooden things), 4 consultants rejected the CP kwe. For Domain 17 (texts), 2 consultants rejected the CP kwe and 5 consultants rejected the CP ke. And for Domain 20 (names), 4 consultants rejected the CP kwe. From a sociolinguistic point of view, this result is not surprising; it is a nice example of the generally observable discrepancy between "official" language norms and actual language production (see also Unterbeck 1990b:60-61). This discrepancy is observed especially when consultants are confronted with such official norm-violating utterances presented out of context. A number of consultants also rejected some CP types as members of the semantic domains proposed. I mention here only those cases where 6 or more of the 24 consultants rejected a certain CP within a certain semantic domain: 7 consultants rejected kudu for Domain 1b (3 of these consultants substituted ka'i for kudu); 9 consultants rejected yam for Domain 1b (2 of these consultants substituted kweya for yam); for Domain 6, 6 consultants each rejected utu, kala, and pila, and 8 consultants rejected meila; for Domain 7,6 consultants each rejected siwa and va and 11 consultants rejected liku. I should note here that 4 consultants offered the CP katupo ('section, quarter'; see section 3.3.1, Table 3.3.2) as a substitute for the CP liku. With the exception of this substitution, all these observations confirm the results
Six Years Later—A Restudy
299
of the evaluation of CP types within their semantic domains presented previously in section 3.3.4.2. Obviously, the rank of these CPs within their semantic domains is so low that they are no longer regarded as members of the domain by a number of native speakers. In addition, consultant 20, Vadomna, emphasized that the CP num ('magic, magical formula') belongs to Domain 17 (texts). This CP, however, is very rarely used and is almost obsolete; only some of my older consultants in Tauwema recognize it. Furthermore, after the completion of this elicitation test, during one of our evening "gossip" sessions with our friends in Tauwema, the CPs sebulu ('grass skirt for little girls') and tili ('bits of lime clinging at a lime spatula') were also produced (see section 3.3.1, Table 3.3.1). Thus, just as katupo should be mentioned as a member of Domain 7 and num as a member of Domain 17, the CP tili should also be mentioned as a member of Domain 3b (quantities, things) and the CP sebulu as a member of Domain 19 (dress, adornment). However, it is obvious that these CP types play only a minor role in Kilivila. Thus, the results of my restudy confirm my grouping of the 88 CP types into the 20 semantic domains presented in section 3.3.4. Kilivila native speakers accept the semantic domains proposed. Their modification of these domains was marginal: on the one hand, they rejected only CPs that rank very low within a domain and, on the other hand, they added only three new CP types in three different domains; these new CPs are very rarely used and are almost obsolete. To answer the question of whether Kilivila native speakers accept, reject, or modify the variable rules formulated in section 3.3.4.2—that is, to check whether Kilivila native speakers producing a certain CP type referring to a certain nominal concept really adhere to the variable rules formulated in the section mentioned—I proceeded as follows: I designed a language evaluation test in which I, as the experimenter, presented my consultants with a number of phrases and then asked them to evaluate them. I approached at random 24 consultants (12 female and 12 male) belonging to four different age groups (see appendix C) and asked them to cooperate. Previous tests showed that it was not useful to include children between 4 and 7 years of age in such a complex test. The number of phrases presented during this test depended on the alternatives that the 87 variable rules, formulated in section 3.3.4.2, provide for speakers. Thus, the number of phrases varied from one to seven (seven phrases occurred only twice); fortunately this number of phrases is within the limits of our capacity for processing information (Miller 1956). The order in which I presented the phrases was random to exclude information processing variables such as the "last in, first out" phenomenon. Moreover, I randomly alternated the word classes that incorporated the CPs; however, checking the tape recorded interviews I realized that I used the CPs most often with demonstrative pronouns. To illustrate my method of data collection, I give the following example using the standardized questions for the variable rule of CP production to refer to the concept 'tooth' (see section 3.3.4.2, Tables 3.4.1.2a and b):
300
Classificatory Particles in Kilivila
[Consultant's name], kulivala "makena kudula" kena "makuduna kudula" kena "maka'ina kudula" kena "makwena kudula"? (Free translation: '[Consultant's name], do you say "this (ke) tooth" or "this (kudu) tooth" or "this (ka'i) tooth" or "this (kwe) tooth?") If the consultant answered that he or she uses all these variants, I specified the consultant's task in the following way: Gala, magigu yokwa kukanta'i kwetala wala bukulivala. (Free translation: 'No, I want you to decide on one [phrase] only and you will say it.') After the consultant decided on a phrase, I asked him or her: Kumwedona avaka alivala biga bwena, kida? (Free translation: 'Everything I said is correct [Kilivila], isn't it?) If this question had been answered negatively, which never happened during the test, I would have wanted the consultant to correct the unacceptable phrase. After this checking of my test phrases, I asked the following two questions and noted the consultants reactions: Amakala kudoki guyau mokita bukulivala! (Free translation: 'What do you think a real big chief would say?') Amakala kudoki tommwaya tokunibogwa elivalisf! (Free translation: 'What do you think the ancestors would have said?') I asked these questions to get more information about sociolinguistic variables and the consultants' linguistic awareness with respect to these variables, on the one hand, and to elicit some archaic CPs, on the other hand. However, this plan failed; all the consultants claimed that the big chiefs as well as the ancestors would have decided on exactly the same phrases they chose. All the noun phrases used during this language evaluation test are given in appendix C. The data gathered during this test were ordered and analyzed in two ways. First, I noted the CPs the 24 consultants chose to refer to the 87 nominal concepts that I presented. I counted the tokens of the CPs and calculated their relative frequencies (r.f.s) for all 24 consultants (if all the consultants produced the same CP type, the r.f. for this CP type is 1.00; if the r.f.s for the CPs produced with one variable rule do not add up to 1.00, this means that the consultants could not decide on one CP or did not know the particular CP type or types). I then compared these r.f.s with the r.f.s I computed for the CP production in the 1983 database (see r.f.s in Tables 3.4.1.la to 3.4.20a in section 3.3.4.2). Second, I checked the consultant's choices and compared them with the variable rules of CP production formulated in section 3.3.4.2. These variable rules predict the consultants' CP choice when they want to refer to nominal concepts. I counted how many of my consultant's responses were the predicted first or second choice (according to the r.f.s given in the variable rules). I also counted the number of CPs that function as sociolinguistic variables (SLVs) and checked the choice of SLVs with the consultants' clan membership. I then calculated the percentage of my 24 consultants' speech behavior that was correctly predicted by the variable rules formulated in section 3.3.4.2 Table 4.1 summarizes the results of the first analysis of these data. This table reflects the following
Six Years Later—A Restudy
301
Table 4.1. Relative frequencies of CPs within the 20 semantic domains in the restudy compared with those in the 1983 database Semantic domain la
1b
Nominal concept male person female person child all humans all humans corpse spirit limb hand
tooth
2
3a
animal shell group/team group move
shoal
batch/fish
fish/strings
3b
heap
plantation
CP to na gudi tolna/gudi to na na kweya kwe ke kweya to kwe yam ya ka'i kudu ke kwe na kwe buda kwe deli kasa to ke kwe na buda kwe kasa yuva ke suya wela na wela na oyla gula kwe tam buda kwe po'ula buda gula
r.f. (restudy)
r.f. (1983 database) 7.00 .96 .79
.83 .17 .96 .96 .96 .04 — .92 .08 — — — 1.00 — — — 1.00 .96 .71 .29 .50 .38 .04 .04 — — .67 .25 .08 — — .46 .46 .08 .50 .46 .04 .50 .29 .17 .04 .29 .46 .08 —
— — .73 .05 .03 .47 .02 .31 .07 .01 .69 .14 .01 .01 .90 — .35 .47 .52 .01 .06 — .21 .01 .48 .29 .01 .02 .01 .50 .07 .33 .17 .29 — .28 .46 .16 .01 .52 .27 — .02
Classificatory Particles in Kilivila
302
Table 4.1. (continued) Semantic domain
Nominal concept fruit cluster
CP ke po'ula luva bukwa
ya
cluster/bundle
bundle/wrapped small bit
bit/piece
part/piece
scrap
cluster/bananas conical bundle
nut bunch
luba kwe sa duli kwe ke luba kapwa kwe utu ke si bwa vili kwe kasa gum kwe ke utu buda ya pila utu ya ke kwe vili bubwa utu bwa ke ke kila kwe mmwa luva ke guba kwe buda sa po'ula luva ke kwe
r.f. (restudy)
r.f. (1983 database)
.37 .21 .21 .13 .04 .04 — — .83 .17 — .67 .25 .08 .46 .29 .13 .13 — — .37 .29 .13 .13 .04 .04
.68 .01 .01 .02 .08 — .02 .02 .60 .16 .03 .43 .26 .17 .66 .16 .01 .01 .02 .02 .02 .49 .12 .03 .01
—
.50 .33 .13 .04 — — .37 .33 .17 .13 .54 .46 — .88 .08 .04 — — — .46 .33 .17 .04 —
—
.03 .45 .01 .33 .03 .02 .01 .04 .48 .11 .20 .79 .04 .04 .13 .02 .67 .02 .02 .01 .13 .02 .02 .63 .11
Six Years Later—A Restudy
303
Table 4.1. (continued) Semantic domain
Nominal concept sheaf
bundles of four things
handful
CP
r.f.
r.f. (1983
(restudy)
database)
sipu kwe ya utu ke kwe yulai ke utu ka'i yeni
.71
.13 .08 .04 .04 .71 .13 .08 .04 .04 1.00
.49 .27 .04
1.00 1.00
4
anything
kwe ke
5
span
ke uva tam kwe yuma ta kwe bogi kwe tuta yam kala kwe tuta siva utu pila meila kwe kabulo ke utu
length 6
night
day
time
day/part 7
village sector
place/area sea portion land plot
kwe kwe vilo
utu kwe utu bwalita kubila kwe
ke pila corner/garden
kwe nunu kabulo utu
.50 .25 .21 .04 .58 .08 .71 .29 .79 .13 .04 .04 .71 .17 .13 — — .63 .08 .75 .13 .08 .04 .79 .17 .04 .88 .13 — .83 .17 — .63 .25 .08 .04
—
.01 .82 .05 .01 — — — .91 — .61 .17 .04 .04
— —
.77 .07 .80 .01 .03 .02 .74 .04 .02 .03 .01
— _
.12 .51 .01 .22 .85 — — .76 .02 .03 .52 .29 .01 .01 .69 .12 .02 .01
304
Classificatory Particles in Kilivila
Table 4.1. (continued) Semantic domain
Nominal concept mountain swamp moon star area/authority river/creek
8
cut
punctured
cut off
cut across
flexible/thin
untwisted
9
protuberances
r.f.
r.f. (1983
(restudy)
database)
kwe koya dumia kwe na na kwe liku kabulo va/vaya kwe
.92 .08 .71 .17 1.00 .96 .04 .88 .04 .88 .04
.73 .14 — — — — — — — — —
giwi kwe ke kasa pwanina/pona nigwa kwe bubwa ke utu bwa kwe bubwa ke bubu kwe utu ya bwa sisi ya ke pila si ke vili utu bwa kwe
.88 .08 .04 — .75 .25 — .50 .29 .17 .04 — .42 .21 .17 .13 .04 .04 — .75 .13 .08 .04 — .50 .37 .08 .04 — .54 .25 .17 .04 1.00 — — .37 .33 .08
.59 .06 .09 .02 .78 .01 .13 .09 .72 .01 .09 .01 .01 .10 .38 .16
CP
ke utu kabulo
row/line
row
bwa kasa kwe ke gili
ya vakala
—
.02 .02 .61 .07 .10 .02 .08 .27 .38 .17 .01 .02 .59 .21 — .02
.77 .05 .01 .07 .56 .01
305
Six Years Later—A Restudy Table 4.1. (continued) Semantic domain
Nominal concept
kneaded/dot
10
wood trees bough/cut off bough/branch wooden dishes
carving/human
11
r.f. (1983 database)
kwe vili utu kasa notu kwe
.08 .04 .04 — .50 .50 .63 .37 .54 .46 .92 .08 .83 .17
.16 .02 — .02 .57 .18 .92 .01 .03 .86 .78 .03 .84 — .83 .02 .01 — .01 .04 .03 .02 .19 — .08 — .01 .88 .01 .01 .53 .27 .03 — .19 .65
ke bwa bwa ke sisi ya ke lila
ke luva kwe na
fish trap yams/basketful
tam
basket pot/potlike
13
r.f. (restudy)
pila ke kwe kavi ke oyla kwe wela pila to kwe ya kwoila kwe ke kauya
tool
string
12
CP
compartment/ foodhouse
compartment/canoe/ foodhouse compartment/creel
1.00
— —
.96 .46 .29 .17 .08 .29 .29 .21 .13 .08 .96 .04 — .88 .13
—
kwe 0 tetu
.96 .79 .21 — —
kabisi ke bwa kwe utu pila
.54 .33 .08 .04 — —
.43 .29 .02 .06 .02 .02
liku ke kwe lipu
.54 .42 .04
.39 .18 .31
.88
-
.03
.01
306
Classificatory Particles in Kilivila
Table 4.1. (continued) Semantic domain
Nominal concept
14
door/entrance
15
fire/fireplace earth oven
16
road/track walk/journey
trip paddle strike
17
part/song
part/song/magic
division/magic
CP
stone blade ginger
.29 .17 .17 .13 .08 .08 .04 .04 .88 .08 .04 1. 00 —
kada ya kwe kada lilo ke bililo/bili vili duli kali/kai pila kasa kwe nina meila megwa utu pila meila kasa utu vosi ke kwe sisi megwa
1.00 .46 .33 .13 .08 — .88 .04 .04 .96
kai beku kwe sam tam luva
19
band of fibers
r.f. (1983 database)
ke kada duya kwe ya pila va tabuda kova ke kwe kumla kwe
kwe utu
18
r.f. (restudy)
kwe kudu
.71 .13 .08 .04 .04
— —
.50 .21 .17 .13 — — — .79 .04
.04 .04 .71 .21 .08 .71 .17 .08 .58
.33
.07 .51 .05 .07 .12
—
.05 — .88 .01 .01 .90 .01
.84 .09 .73 .01 .04 .01 — — — — .37 .02 .34 .02
—
.07 .02 .51 .02 .06 .02 .10 .07 .05 — — — — .76 .01 .01 — — .81
.03
Six Years Later—A Restudy
307
Table 4.1. (continued) Semantic domain
Nominal concept
belt/shells
20
name
CP utu luba doba vakala gili suya kwe iga/yegila kwe
r.f. (restudy)
r.f. (1983 database)
.04 .04 — .83 .08 .04 .04 .96 .04
--
— .01
— — — —
.83 .03
trends in the consultants' reactions to the language evaluation test presented. First, we notice that the presentation of CP variants with which a speaker can refer to a certain nominal concept in isolation from context (a presentation similar to that of minimal pairs) induces the consultants to choose special and more specifying CP types over the more general ones offered by the variable rule as reference variants (see, e.g., Domain 3a, 'fish/strings'; Domain 3b, 'cluster/bundle', 'cluster/bananas', 'nut bunch'; Domain 7, 'village sector'; Domain 8, 'punctured'; Domain 11, 'pot'; Domain 17, 'part/song/magic'). Moreover, with the CP variants presented in this way, the consultants obviously prefer to choose CPs that function as sociolinguistic variables, even if they are not persons of high status within their speech community (see, e.g., Domain 3a, 'batch/fish'—wela, 'fish/strings'—wela; Domain 3b, 'scrap'—bubwa, 'cluster/bananas'—kila, 'conical bundle'—mmwa, 'nut bunch'—sa; Domain 7, 'corner/garden'—nunu; Domain 8, 'cut across'—bubwa, 'cut off—bubwa; Domain 9, 'row'—gili; Domain 10, 'wood'—bwa, 'tree'—bwa; Domain 13, 'compartment/canoe/foodhouse'—liku). However, this is not surprising, either, from the sociolinguistic point of view, if data like these are presented in a way that is similar to the presentation of minimal pairs. Second, the language evaluation test provides us with interesting information about some CPs that were mentioned in connection with the variable rules within certain semantic domains but could not be evaluated with respect to their status within these domains because of lack of sufficient data (see, e.g., Domain 7, 'village'—CP vilo; Domain 9, 'protuberances'—CP kabulo; Domain 13, 'compartment/creel'—CP lipu; Domain 16, 'paddle strike'—CP kali; Domain 17, 'division/magic'—CP sisi). Third, the language evaluation test provides us with some new information: Thus, consultants 22 and 24 (see appendix C) produced the CP ta to refer to the concept 'span' in Domain 5 instead of the CPs presented for evaluation. This usage can be explained by the fact that a palm leaf as big as the span of two extended arms is generally used to weave a basket (remember that the CP ta refers to the concept 'basket'). Therefore, we must consider the CP ta, as a member of Domain 5 (measures).
308
Classificatory Particles in Kilivila
Consultants 15 and 17 used the CP wela 'string of fish' to refer to the concept 'string' in Domain 11 (utensils) instead of the other CPs presented for evaluation; obviously, the CP wela loses its more specific meaning in this case. Therefore, wela must be considered a member of Domain 11. For Domain 14 consultant 8 produced the CP tabuda to refer to the concept 'door, entrance' instead of the other CPs presented for evaluation. Tabuda is obviously a variant of the CP tabudo (Lawton 1980; see Table 3.3.2 in section 3.3.1). I conclude that this CP also has the meaning 'room' (which Lawton mentions) and the meaning 'door entrance'. Therefore, the CP tabuda/'tabudo must be considered a member of Domain 14. The consultants' reactions to the phrase presented to check the use of the CP sam in Domain 18 (ritual items) is extremely interesting. Most of the consultants explained that the CP sam is only used in connection with play accompanying verses to refer to the concept 'ginger'; in all other contexts, the CPs tam 'sprouting' and luva '(tied) bundle' should be used to refer to this concept. Therefore, these two CP types must be considered members of Domain 18, too. Finally, in connection with Domain 19 (dress, adornment), the consultants used the CPs utu and luba to refer to the concept 'band of fibers' and the CPs gili and suya to refer to the concept 'belt' instead of the other CPs presented for evaluation. Therefore, these four CPs must be considered members of Domain 19, too. Fourth, this evaluation test modified a few of the results presented in section 3.3.4.2. Thus, the use of the CP kasa in Domain 3b to refer to the concept 'bit, piece' is obviously not as idiosyncratic as claimed, and the CP pila, used in Domain 11 to refer to the concept 'tool', is obviously much more important for this domain than was claimed on the basis of the 1983 data. Whether these modifications represent language change, however, is a question that must remain open. Finally, this first ordering and analysis of the evaluation test data not only confirms some of the trends observed with respect to processes of language change (processes that predicted that the CP ka'i is going to supersede the CP kudu in Domain 1b and that the CP kwe is going to supersede the CP koya in Domain 7, to give just the two most striking examples), but also confirms the results presented in section 3.3.4.2 in general. After this first data analysis, I ordered and analyzed the data so that I could check the consultants' CP choices in the evaluation test in relation to the variable rules formulated in section 3.3.4.2 As mentioned above, these variable rules predict the consultants' CP choice when they want to refer to certain nominal concepts. To test the predictive power of these variable rules, I counted my consultants responses with respect to the first and second choice to be expected according to the relative frequencies as indicators of CP type production given in these variable rules. Moreover, as mentioned, I also counted the number of CPs that function as sociolinguistic variables (SLVs) that were chosen by the consultants during the language evaluation test and recorded the clan membership of the consultants producing these SLVs. Table 4.2 presents the results of this analysis.
Table 4.2. The predictive power of the variable rules in 24 consultants' responses to the phrases presented in the CP language evaluation test
Consultant no. (age, years)
1
Clan membership
No. of 1st choice expected in 87 rules
No. of 2nd choice expected in 87 rules 13 13
2 3 4 5 6
(12) (10) (13) (10) (12) (14)
Malasi Malasi Malasi Malasi Lukwasisiga Lukwasisiga
49 43 61 50 54
7 8 9 10 11 12
(17) (20) (19) (16) (15) (15)
Malasi Lukwasisiga Malasi Malasi Malasi Malasi
54 58 49
13 14 15 16 17 18
(35) (25) (30) (26) (30) (34)
19 20 21 22 23 24
(39) (48) (54) (45) (62) (69)
Malasi Malasi Malasi Lukwasisiga Malasi Malasi Lukwasisiga Malasi Malasi Lukwasisiga Malasi Malasi
63
58 72
64 63 51 62 61 61 54 64
55 65 61 69 63
7 7 14 10
7 6 10 6 2 7
8 8 8 6 9 10 8 7 6 10
No. Of SLV CPs chosen 7
6 10 8 6 2
9 5
5 7
7 11 9 7 6 7
8 10 5
4
9 11 6 12
9
7
Percent CP production correctly predicted (100% = 87 rules) 79 71 90 75 78 (85t) 84 (86t) 81 74 (79t)
74 82 93 94 92 76 87
77 (85t)
90 85 89 82 94 82 (89t)
98 91 (continued on next page)
Table 4.2. (continued) Consultant nos.
1-3 7-9 13-15 19 21 1 3 7-9 13-15 19-21
Mean value (%)
80 76(78 ) 85 88
82 (83t)
Consultant nos.
4-6 10-12 16-18 22-24 4-6 10-12 16-18 22-24
SLV — sociolinguistic variable. SLV included, although consultant is not a person of high social status with respect to clan membership.
Mean value (%)
79 (82t) 90 84(87
)
90 (93t)
86 (88t)
Six Years Later—A Restudy
311
As Table 4.2 documents, the range of correctly predicted CP types that Kilivila native speakers use to refer to certain nominal concepts varies from 71% (consultant 2) to 98% (consultant 23). If we look at the mean values, the variable rules correctly predicted 82% (83% when SLV is added) of the 12 female consultants' CP production (with a range of 71% to 94%) and 86% (88% when SLV is added) of the 12 male consultants' CP production (with a range of 75% to 98%). Although some of the consultants choosing CPs that function as SLVs do not belong to the group of persons of status defined on the basis of their clan membership, we can nevertheless consider these cases of CP production in our analysis, keeping in mind the general finding of sociolinguistics that with speech data presented in isolation most speakers chose the most prestigious variant, at least in a test situation, even if (as in the present case) they would face sanctions like ridicule, e.g., if they used these variants in their own public language production. Anyhow, with all due modesty but also with the field researcher's legitimate pride after the successful confrontation of his theoretical analyses with empirical reality I can state that the variable rules formulated in section 3.3.4.2 describing Kilivila native speakers' CP production are verified, confirmed, and proved to be accurate on the basis of this restudy in a rather impressive way.
Chapter
5 Excursus: Language, Culture, and Cognition? . , . even if there were a large number of roots for different snow types in some Arctic language, this would not, objectively, be intellectually interesting; it would be a most mundane and unremarkable fact. Horse breeders have various names for breeds, sizes, and ages of horses; botanists have names for leaf shapes; interior decorators have names for shades of mauve; printers have many different names for different fonts (Caslon, Garamond, Helvetica, Times Roman, and so on), naturally enough. If these obvious truths of specialization are supposed to be interesting facts about language, thought, and culture, then I'm sorry, but include me out. GEOFFREY K. PULLUM (1991:165) In this chapter, I will discuss possible interdependencies between language, culture, and cognition as they relate to the CP system in Kilivila. This is a venture that requires great caution. I am certaintly aware that most of what will be said in this chapter is highly speculative; nevertheless, I dare to set out on the rather uncertain ground of speculation recalling Albert Szent-Gyorgyi's statement, quoted by Koestler (1978:154): "There is but one safe way to avoid mistakes: to do nothing ...." The discussion of human categorization has a long philosophical tradition. Royen (1929:1), for example, emphasizes that even the linguistic problem this volume addresses, namely, the question of nominal classification, can be traced back to the Greek sophistic philosopher Protagoras (485-414 B.C.). Obviously, the problems of "category" and "categorization" and especially the interdependencies between category, categorization, and classification, on the one hand, and naming, language, thought, perception, and culture, on the other hand, have been discussed as an important feature of human 312
Excursus: Language, Culture, and Cognition?
313
existence at least since the ancient Greek schools of philosophy. We may even quote the famous first lines of the evangelist John's gospel: In principio erat verbum, et verbum erat apud Deum, et Deus erat verbum. Hoc erat in principio apud Deum. Omnia per ipsum facta sunt: et sine ipso factum est nihil, quod factum est, in ipso vita erat, wt vita erat lux hominum: et lux in tenebris lucet et tenebrae earn non comprehenderunt. Thus, we may argue that this philosophical problem is an original one, indeed. The history of the discussion of this topic is fascinating (see, e.g., Foucault 1966; Rosch 1988; Vollmer 1988a and b); however, in this excursus I can only emphasize once again that in linguistics the study of the question of the interdependence between categorization, classification, language, thought, perception, mind, and culture has a long tradition as well. The work of Herder (1770), Humboldt (1836), (see also Heeschen 1972; Malmberg 1987), Schleiermacher (1838), and Whorf (1958) is familiar to most linguists, I suppose, and still provokes fervent discussions about the sense and nonsense of the so-called linguistic relativity principle or the Sapir-Whorf hypothesis, which Whorf (1958:5) summarized as follows: We are thus introduced to a new principle of relativity, which holds that all observers are not led by the same physical evidence to the same picture of the universe, unless their linguistic backgrounds are similar, or can in some ways be calibrated. In 1972, Gipper published an excellent monograph summarizing the discussion of this hypothesis, and, more recently, the contributors to Noun Classes and Categorization edited by Colette Craig (1986), especially Lakoff, have taken up this discussion again.1 Lakoff (1987) carries on the debate using Rosch's pioneering research on the problem (see Rosch 1978; 1988; Rosch and Lloyd 1978) as his main point of departure to present extremely interesting insights. He provides some evidence against the traditional "objectivist view" and for the new "experientalist view" of human reasoning, claiming that reason is embodied and imaginative. Throughout this book it is not only emphasized but also quite evident that a discussion of the interdependencies among categorization, classification, language, thought, 1 In connection with the general problem, as well as with its relevance for classifiers, see also Durkheim and Mauss (1970:vii, 9, 81; Broadfield 1946:1, 25, 27; Lenneberg 1953; Carroll and Casagrande 1958: 18-19; Bruner et al. 1962:245-255, 247, 277, 311-312; Frake 1969:30; Tyler 1969:3, 6; Cole et al. 1971: Leach 1972:43-44; Berry and Dasen 1974; Cole and Scribner 1974; Haugen 1977; Lenneberg 1977:401-11, 444 448; Denny 1979:102; Ellen 1979:3; Hallpikc 1979:2, 25, 79, 92: Harweg 1987b:264; Bechert 1988:1.
314
Classificatory Particles in Kilivila
perception, and culture is necessarily an interdisciplinary enterprise involving not only linguists, but also anthropologists, psychologists, and philosophers. This insight is not, strictly speaking, original; the issue was also raised in the work of Malinowski (1922:1-25), Royen (1929:iii, 37, 85, 98, 110, 192, 206, 269, 363-364, 889), Schieffelin and Ochs (1986:164-165, 168-169, 183), and Watson-Gegeo and Gegeo (1986), to name just a few; however, the present situation seems to require this emphasis, and we can only regret that insights that were taken for granted in the time of Wilhelm Wundt and Karl Buhler need special reference today. In any event, the following paragraphs discuss what kinds of interdependencies among language, perception, culture, and cognition can be detected if we look at the system of CPs in Kilivila. As a starting point for our discussion, we should keep in mind the literature just mentioned, as well as Lakoffs statement that there "is nothing more basic than categorization to our thought, action, and speech" (Lakoff 1987:5) and Rosch's comment that when "we speak of the formation of categories we mean their formation in the culture" (Rosch 1978:28). Moreover, a speaker of any classifier language, and especially a speaker of Kilivila, must classify all nominal denotata—an infinite set probably—with CPs that may, in theory, be infinite but in everyday speech constitute a finite set of formatives; thus, the statements that "classifiers are linguistic correlates to perception" (Allan 1977:308) and "linguistic classifiers relate people to the world' (Becker 1975:118) are quite plausible and convincing (see also Unterbeck 1990b:43). One of the main functions of classification and categorization is to organize and order the perceived world. Although human cognition is a result of biological evolution and thus is constrained by a number of biological filters, as psychologists and adherents of the philosophy of evolutionary epistemology have shown (see, e.g., Vollmer 1988a and b; Koestler 1978:201), nevertheless we need to further "cut down the diversity of objects and events" with which we as "organism(s) of limited capacities" must deal (see Bruner et al. 1962:245). Otherwise, our cognitive capacities would be stretched beyond their limits. Thus, we may interpret the semantic domains constituted by the system of Kilivila CPs as categories that native speakers have developed (and are still developing) to order their perceived world, as it is encoded and represented in the nominal denotata of their language. This interpretation assigns to the semantic domains constituted by the CPs the status of linguistic manifestations of Trobriand classification and categorization of their perceived world. The questions to be raised and answered now are the following (see G. Senft 1989:162): Do the linguistic manifestations of the Trobriand perception of the world allow any kind of inferences to Trobriand cognition? Do these categories "frame" Trobriand thought, in Goffman's (1974) sense? Do the linguistic manifestations of the Trobriand perception of the world allow any kind of inferences to Trobriand culture? Or, alternatively, Do these linguistic manifestations of Trobriand perception represent universals of human cognitive processes, completely or in part, or do they merely represent language-specific or culture-specific characteristics of Trobriand thought? These questions make clear the speculative nature of the arguments that
Excursus: Language, Culture, and Cognition?
315
will be used to attempt to answer these questions. In the following paragraphs, I rely heavily on the hierarchical order of the semantic domains constituted by the CPs, as presented in Table 3.4.22 in chapter 3. Of course, this hierarchy is itself based on speculation, as stated in section 3.3.4.2. Nevertheless, I hope that we will be rewarded for this daring enterprise by some further ideas and maybe even some insights. In chapter 3, I grouped the 88 CP types into 20 semantic domains with two domains containing subdomains. This grouping was primarily based on commonsense considerations; however, as the restudy presented in chapter 4 showed, this grouping is not at all an etic one; on the contrary, my consultants accepted the semantic domains proposed, modifying them only marginally. I will now discuss the 20 semantic domains in the hierarchical order presented in Table 3.4.22 with respect to the questions raised here. 1. Persons If we consider Domain la as a subdomain of Domain 1, the semantic domain covering the concept 'persons' is on top of the hierarchy and thus is the most important concept for linguistic reference. This result is not extraordinary. As developmental psychologists and human ethologists have documented for many cultures, and as we all know, to refer to, and to communicate with, a fellow human is a strong and universal need for Man, the ens sociale. This is also reflected in the fact that CPs in Domain la are the first to be acquired by Trobriand children. Moreover, if we look at classifier languages in general, the semantic domain for humans seems to have a special status. For example, Adams (1989:177) points out that classes "for humans, because they represent or include the self, are unlikely candidates for oblivion" (see also Adams 1992). However, the Tzotzil dialects Mitontik and Chenalho provide an interesting counterexample, because they no longer have a classifier for persons (de Leon 1988:108). 2. General CPs It is not surprising that the domain containing the two general CPs comes next in the hierarchy of semantic domains. As pointed out previously, the general CPs can be used to refer to everything except human beings; they are acquired at the same time as the CPs that refer to human beings, that is, in the first phase of CP acquisition. They allow the production of grammatically correct noun phrases with numerals, demonstrative pronouns, and adjectives, and they are produced in connection with most abstract nouns and most loan words (see G. Senft 1991a: 1992a). Adams (1989:177) comments on the important role of general CPs with respect to processes of language change, a role that was discussed in detail in section 3.3.3. Adams refers to general classifiers as "cannibalising classifiers" (because they devour other classifiers in the system). She mentions the example of Khasi, a Mon-Khmer language that "has only two classifiers, one for humans and one for everything else," thus emphasizing the special status of CPs belonging to the domains 'persons' and 'general CPs'.
316
Classificatory Particles in Kilivila
3. Animals The semantic domain 'animals' comes third in our hierarchy of semantic domains, which is understandable because animals play a rather important role in Trobriand life. Fishing is one of the two main sources (the other is gardening) of the Trobrianders' subsistence (see Malinowski 1935; BellKrannhals 1987). My data document the Trobrianders' incredible abilities of taxonomic differentation with respect to fish and shells (clams and snails). With respect to gardening, the Trobrianders have to safeguard their gardens against roaming wild bush pigs, bush kangaroos, and wallabies; the knowledge of a number of insects, especially of vermin, is, of course, very important in gardening. The slaughtering and communal eating of domestic pigs is one of the main events of a number of Trobriand festivals. In addition, animals play a prominent role in Trobriand culture. Malinowski (1974:112) described how the Trobrianders' four main clans are associated with clan animals (the iguana for the Lukulabuta clan, the dog for the Lukuba clan, the pig for the Malasi clan, and the crocodile, snake, or opossum for the Lukwasisiga clan). The myth about the behavior of these animals, which are the clan representatives or totems, after their emergence from a hole close to the village of Labai on Kiriwina island lays the foundation for the ranking of the four main clans, a fact extremely important for the understanding of the Trobriand social construction of reality. In addition, the members of the socially stratified clans must obey a number of food taboos that also affect animals, especially fish (see, e.g., Malinowski 1929:26-27). Thus, there are general as well as culture-specific arguments that explain the importance of this semantic domain for the Trobrianders. 4. Trees, wooden things The group of the most important semantic domains is completed by the domain 'trees, wooden things'. Until quite recently, almost everything of the Trobriand material culture—from the lime spatula to the house and the canoe—was made out of wood and/or leaves of trees or at least had a wooden component.2 Moreover, the overall importance of the coconut palm for human life in Oceania is well known. 5. Persons/body parts The fifth place in our hierarchy of semantic domains constituted by Kilivila CPs is occupied by the compound domain 'person and body part'. However, for the purposes pursued here, it seems to be more appropriate to split up this compound domain and to treat the two subdomains separately in our discussion. 2 To give just one example: During my data collection on how to build a masavva-type canoe and how to make its sail, my consultants named 24 different kinds of trees and other wooden materials that are needed for the construction of this impressive boat (see also Malinowski: 1922, chap IV and V).
Excursus: Language, Culture, and Cognition?
317
6. Place Orientation abilities are necessary for our survival as a species (Schone 1983). In some situations, we are very much aware of this fact, whereas in other situations we seem to be completely unconscious of the working mechanisms of these orientation abilities (see Hallpike 1979:304-313). Given the general necessity of methods for deictic reference for survival, it is not surprising that the domain "place" belongs to the second group of important semantic domains constituted by Kilivila CPs. 7. Quantities For our purposes, it seems appropriate to split this compound domain into the two subdomains 'quantities, with generally inanimate referents' and 'quantities, with generally animate referents' and to treat these two subdomains separately in our discussion. 8. Quantities, with generally inanimate referents Reference to quantities of things is a rather basic need in human communication. The Trobrianders are highly competitive (see Malinowski 1935, vol. 1: chaps. V and VI; Powell 1957, 1960; Hutchins 1980; G. Senft 1987d: 195-203; Weiner 1988), and they test the socially regulative norms of behavior in everyday ritualized communicative forms of requesting, giving, and taking, for example, tobacco or betel nuts (see G. Senft 1987c: 109-111). Therefore, the domain 'quantities, with generally inanimate references' is especially important, and it is easy to understand that this subdomain is another member of the second group of important semantic domains constituted by the CPs. 9. Fire, oven As in European mythology and philosophical tradition, where we have the myth of Prometheus, on the one hand, and that of Heraklit, who claimed that fire was the primary principle of the world, on the other hand (see Vollmer 1989, 202-204), fire also plays an extremely important role in Oceanic mythology. However, these mythological aspects aside, the importance of both a fire and an oven in everyday life and the need to refer to these concepts verbally is obvious. Thus, it might have been expected that this semantic domain would belong to the second group of important semantic domains constituted by the CPs. 10. Names In the Trobrianders' hierarchically structured clan society, proper names are clan property.3 Thus, anyone hearing the name of another person can 3 In the Trobriands, we find a matrilineal society with patrilocal residence. Children belong to their mother's clan. This has far-reaching consequences for Trobriand society in general. Hereditary titles and land rights, as well as myths and magical formulae are obtained through the mother's side (see, e.g., Malinowski 1929; Hutchins 1980; G. Senft 1985c, 1985d; Weiner 1976, 1988).
318
Classificatory Particles in Kilivila
immediately identify their clan membership. This is true for the whole Massim area. If we keep in mind that the Massim4 are famous seafarers and cover long distances with their canoes, and that they stop at various places and are dependent on the hospitality of other islanders in the area, this principle is very important, indeed. Wherever the Massim put their canoe ashore, they are welcomed and asked their names. They are immediately identified as members of certain clans and thus can hope to rely on the hospitality of their co-clan members. Thus, the clans' possession of proper names creates a more or less binding net of kinship throughout the Massim area. Therefore, the concept 'name' is rather important for the Trobrianders, especially on journeys and trips. However, the importance of this concept also holds for other ethnic groups. The semantic domain 'name' is the first of a group of domains with intermediate importance in Kilivila. 11. Time Everyone familiar with the "Melanesian way of life" and especially with the Trobrianders' everyday idea of time must be astonished to find this category so high in our hierarchy of semantic domains. However, if we keep in mind the importance of the Trobriand calendar, especially for the seasonal rhythm of gardening (Malinowski 1935, vol. 1: chap. 1; Austen 1939, 1949/50; Leach 1949/50), this result is not so surprising. This domain holds an intermediate position in the hierarchy of semantic domains constituted by Kilivila CPs. 12. Quantities, with generally animate referents Because the domains covering 'persons' and 'animals' rank rather high in our hierarchy, it may seem surprising that a semantic domain covering quantity, with generally animate referents is in an intermediate position in the domain hierarchy. I do not know how to interpret this fact. 13. Road, journey In discussing the domain 'time', I mentioned the seasonal calendar of the Trobrianders. This calendar is important for gardening, of course; however, as indicated by Malinowski (1935, vol. 1: chap. 1), this calendar affects all the Trobrianders' activities, including journeys and trips. The big "kula" expeditions (see Malinowski: 1922) have their season, too. Therefore, the need to refer to journeys is seasonally constrained and seems to be a matter of secondary importance. Moreover, on islands that are generally easy to survey, the need to refer to roads, most of which are actually small footpaths, does not seem to be vital. It is obviously more important to refer to certain places as a means of orientation. This difference of importance is documented in the rank the domain 'place' occupies within the hierarchy of semantic domains constituted by the CPs; that is, it holds an intermediate rank within the hierarchy. 4
The Trobrianders belong to the North Massim ethnic group. For a geographical definition of the Massim area, see Malinowski (1922:26).
Excursus: Language, Culture, and Cognition?
319
14. Qualities The last domain in the group of semantic domains of intermediate importance covers the concept 'qualities'. Again, the relatively low rank of this domain may be surprising. However, the importance of the domain covering 'trees, wood, wooden things' is such that reference to qualities other than 'being of wooden material' is secondary. 15. Shape The first in the group of not so important semantic domains in our hierarchy covers the concept 'shape'. Considering the emphasis Friedrich (1970) puts on this concept, this is surprising. However, I have no idea how to interpret this finding. 16. Body parts Compared with references to the concept 'person', references to the concept 'body parts' of the comprehensive domain 'persons, body parts' are of minor importance (but see G. Senft 1992c for body part idioms). This is not necessarily true for other ethnic groups in Oceania—for example, consider the different ways of counting that use body parts (see, e.g., Biersack 1982; Koch 1984:122; Lean 1986; Wassmann and Dasen 1994) or of sign languages that use body parts to refer to kinfolk (see, e.g., Kendon 1988: chap. 11). However, in Kilivila, the subdomain 'body parts' belongs to the group of not so important semantic domains within the hierarchy of domains constituted by CPs. 17. Utensils The domain covering 'utensils', like the domain covering 'qualities' seems also to be eclipsed by the domain covering 'trees, wooden things'. Most utensils of the Trobriand material culture are completely or partly made of wooden material. Nevertheless, among the CPs in this domain are those used for referring to baskets (peta) and to clay pots (kwoila), two items that play an important role in Trobriand society (see Malinowski 1922, 1935 vol. 1; G. Senft 1985d; Weiner 1988). However, this domain is obviously of minor importance in the hierarchy of semantic domains constituted by CPs. 18. Dress, adornment Although dress and adornment, especially skirts and body decorations, play an important role in Trobriand culture (see, e.g., Weiner 1976, especially appendix 1; G. Senft in press) the domain that covers this concept occupies a rather low rank within the hierarchy of semantic domains constituted by the CPs. 19. Door, entrance, window Anyone looking at a Trobriand house or a picture of a house will immediately understand why this domain is rather low in the hierarchy of
320
Classificatory Particles in Kilivila
semantic domains constituted by CPs. Obviously, it is sometimes necessary to refer to these concepts; however, this reference is of only minor importance. 20. Ritual items This domain essentially comprises CPs that are used to refer to stone blades (beku). These blades play an important role in Trobriand culture as tokens of wealth (see, e.g., Malinowski 1922: 358, 481; Weiner 1976:179-181, 1988:47). However, they are only very rarely shown and referred to, explaining why this domain ranks so low in the hierarchy of semantic domains. 21. Part of a foodhouse/a canoe/a creel The last domain that belongs to the group of the not so important semantic domains covers the concepts 'part of a foodhouse/a canoe/a creel'. During the filling of the yam houses, on trips in a canoe, and during fishing activities, it is certainly necessary to refer to parts of a foodhouse or a canoe or a creel; however, these references seem to be of only minor importance in Trobriand everyday life. 22. Measures The domain that covers the concept 'measure' is the head of the group that encompasses the least important semantic domains constituted by CPs in Kilivila. The activity of measuring becomes important on the Trobriands when the islanders make a new sail, the size of which must be coordinated with the size of the canoe, and especially after the yam harvest. First the yam tubers are cleaned and piled up in big conical heaps in the gardens. These heaps are critically inspected and measured by the gardener himself, by the fellow gardeners of his village community, and by neighbours who are invited to inspect the harvest (see, e.g., Malinowski 1935, vol 1; illustrations 48, 56, 57, 65; Eibl-Eibesfeldt 1984:401; Schiefenhovel and Bell-Krannhals 1986:29-30, especially illustrations 1 & 3; Weiner 1988:82). Then the long kuvi-type yams are cleaned, fastened to the kedai stakes with which they are transported and displayed, and are carefully measured (see Malinowski 1935, vol. 1; illustration 69; G. Senft 1986:435; Weiner 1988:90). These kuvi yams play an important ritual role on the Trobriand Islands in connection with the buritila'ulo competition (Malinowski 1935, vol 1: chap. V). It is necessary and very important for the Trobrianders to measure sails for their canoes and yams after the harvest; however, the activity of measuring is of secondary importance in their everyday life. Therefore, it is understandable that the semantic domain covering the concept 'measure' is one of the least important domains in the hierarchy of semantic domains constituted by the CPs. 23. Yams As mentioned at the end of section 3.3.4.2, the low rank of the semantic domain that covers the concept 'yams' is probably an artifact of the proposed
Excursus: Language, Culture, and Cognition?
321
computation of the hierarchy index on which we based the hierarchical order of the semantic domains constituted by the CPs. Those familiar with Trobriand ethnography may be doubtful that this domain holds the next to the last rank within the semantic domain hierarchy. Of course, yams are one of the most constitutive factors of Trobriand society (see Malinowski 1935; Weiner 1976; 1988; G. Senft 1985d:476, 490; 1987d). However, when I discussed Domain 12 in section 3.3.4.2, I noted that the zero classifier may be too closely linked with ritual and ceremony and that it is just not produced in profane situations (and thus cannot be elicited in a language production test). This may explain the low rank of this domain within our hierarchy. I am convinced that this domain actually must belong to the first group of the most important semantic domains. 24. Texts The semantic domain that covers the concept 'texts' ranks lowest in our hierarchy of semantic domains constituted by CPs. If we keep in mind that we can expect the knowledge of, and competence to produce, metalinguistic terms than refer to texts and text structure only in expert storytellers, dance masters who coordinate song and dance in festivals, and magicians (see G. Senft 1985c, 1985e), this result is not surprising at all. Four of the nine CPs that constitute this semantic domain are classified in section 3.3.3.1 (see Table 3.3.6) as either CP types that play a secondary role in actual speech production (vosi) or as CP types that belong to a technical language variety or that are extremely rare or almost obsolete (megwa, nina, and meila). I am convinced that this low rank of the domain covering the concept 'texts' also documents processes of fundamental cultural change that have taken place on the Trobriand Islands in the last sixty years or so. Now that we have looked at each of the semantic domains separately, we can try to answer the questions raised earlier in this chapter. The discussion of the domains constituted by the CPs show that, perhaps with the exception of the two domains covering the concepts 'part of a foodhouse/a canoe/a creel' and 'yams', the concepts incorporated in these domains are quite general and seem to be universal for human speech communities. If we alter the label of Domain 23 from 'yams' into 'food' and if we change the label of Domain 21 from 'part of a foodhouse/a canoe/a creel' into 'part of objects' (changes that are quite plausible), this statement holds for all the domains discussed. However, the discussion of the semantic domains also shows that these probably universal categories are defined in a culture-specific way—a result that becomes even more evident if we look at the CPs that actually constitute these domains. This finding fits perfectly well with the general picture we obtain from looking at various classificatory systems (see section 1.3). As the Kilivila CP system shows, the hierarchical order and the culture-specific definitions or "fillings" of these probably universal semantic domains (or categories or concepts) give us much information about speakers' culture
322
Classificatory Particles in Kilivila
and certainly "frame," in Goffman's (1974) sense, the speakers' perception, their kind of perceptive awareness, and their preferred ways of thinking, at least to a certain extent.5 However, this does not imply that this frame cannot be broken or changed if the speech community feels the need to do so. That this changing of frames is generally an "evolutionary" process covering a relatively long period of time is a different matter, indeed; however, it must be kept in mind if we do not want to get trapped in the behaviorists' pitfall once more.
5 See also Slobin (1991:23) who concludes that "we can only talk and understand one another in terms of a particular language. The language or languages that we learn in childhood are not neutral coding systems of objective reality. Rather, each one is a subjective orientation to the world of human experience, and this orientation affects the ways in which we think while we are speaking".
Chapter
6 Closing Remarks: Using Network Models to Describe Classifier Systems In this chapter, I summarize the aims of my study and then present network models for the description of complex classifier systems, based on the insights attained during the analyses of the Kilivila CP system. One of the aims of the study presented in this volume was to list the formatives that constitute the Kilivila CP inventory. The elicitation work I did with my consultants in Tauwema village resulted in the description of the meanings of 91 different CP types; Lawton (1980) lists another 85 CP types,1 and Capell (1969) lists one more CP type that was accepted by my consultants. Therefore, Kilivila makes use of an inventory of at least 177 CPs (see section 3.3.1). My study describes in detail 88 of these CP types. Moreover, for the first time, this study provides information about the actual occurrence of 41 different CP types realized by 1,564 tokens in a corpus of speech data encompassing 34,955 words (see section 3.3.1). Another aim of this study was to determine how this complex morphological system is acquired by Trobriand children; this issue is addressed in sections 3.3.2 and 3.3.4.2. It became obvious that some aspects of this complex psycholinguistic problem could only be understood using information about the actual production of CPs, changes affecting the CP system, and the semantics of the system. The study describes and explains in detail the general temporal progress of the CP acquisition process, the differences in the acquisition process with respect to the production of three of the word classes that use CPs as morphemes in their word formation, 1When I checked Lawton's CP list with my informants in 1989, there were quite a few CPs that they did not know. However, they emphasized that these CPs, as well as CPs that they know but do not themselves produce, are either characteristic markers for the biga galagoki dialect of the people living in Kavataria village (see G. Senft 1986:6) or archaic classifiers. As mentioned in section 3.3.4, 1 do not have references to the CPs Lawton provides in his 1980 thesis in my data (with one exception, mentioned in chapter 4). Thus, I agree with Wittgenstein's maxim: "Wovon man nich sprechen kann, daruber soll man schweigen" (Wittgenstein 1971:115).
323
324
Classificatory Particles in Kilivila
gender-specific differences in the CP acquisition process, the order in which the CPs are acquired, and the reasons that CPs are acquired in a certain order. Moreover, the question of possible parallels between the acquisition process described for Kilivila and that of other classifier languages, as well as English and German, is briefly discussed. The results of the study are too complex to summarize here; I refer the reader to the sections previously mentioned. The third aim of this study was to answer the question of how the CP types are realized in actual speech. Again, the data analyses showed that this complex question, (which seems at first to be a psycholinguistic one) could only be answered in a satisfying way by using sociolinguistic information. The results are too complex to summarize here; they are given in section 3.3.3. The fourth aim of this study was to describe the semantic domains constituted by the CP system in a way that would elucidate the dynamics inherent in the system. On the basis of this description, variable rules of CP production were formulated to predict a speaker's choice of CPs to refer to certain nominal concepts. This description and the variable rules are presented in sections 3.3.4.1 and 3.3.4.2. Before I address this aspect of the study in more detail, I want to mention that chapter 5—an excursus on the general argument—discussed the question of interdependencies among language, culture, and cognition. Rather than summarize the results of this discussion here, I refer the interested reader to the last paragraph of the chapter 5. However, I want to make the following general, mostly methodological remarks. First, the study presented in this volume not only provides the reader with all the methodological devices used to elicit the data, but also presents all the data used in the study. This large-scale presentation requires space, of course; however, it allows the critical reader to check (and countercheck) all the analyses and all the conclusions drawn from these data. Even the restudy is presented in such a way that the analyses can be reproduced. This replicability of results is one of the basic methodological requirements the studies presented here meet without any restrictions. In addition, the fact that the results of the restudy conducted in 1989 confirmed the results of the main study, based on data gathered in 1982 and 1983, emphasizes the extremely high degree of validity of my analyses. Finally, the methods of research presented in this book show that differentiation of linguistics into psycholinguistics, sociolinguistics, ethnolinguistics, and other hyphenated subdisciplines of linguistics is academic and rather arbitrary. Without a sound morphological, grammatical description of the CP system, without information about Trobriand ethnography, without the insights and analyses provided by psycholinguistic as well as sociolinguistic research interests, without any insight into the semantics of the Kilivila CP system and into the pragmatics of the Kilivila speech community—in effect, without the productive and fruitful interplay of all these different disciplines—it would have been impossible to answer the questions raised by this fascinating phenomenon in the Kilivila language.
Using Network Models to Describe Classifier Systems
325
Complex phenomena require comprehensive analyses. Linguists sitting at their desks developing elegant and beautiful theories on the basis of examples taken from their own language competence can avoid the inclusion of these different aspects of their discipline, they are accountable to their peer group of linguists for their own personal dialect only. Linguists doing empirical field research, however, are in a different position. They are accountable to the speech community for their analyses of its language. Their analyses must be as complete as possible, because they are finally verified or falsified by the speech community. It is an empirical fact that the speech community will not accept rearguard action and immunizing strategies with which linguists can withdraw from their analyses to a fuzzy concept such as their "dialect". In the field, the linguist's "dialect" serves only one function: it is a welcome source for their consultants and friends to ridicule them. I now return to the discussion of the semantic system established by the Kilivila CPs. Among other things, this discussion will emphasize once again that linguists doing field research should not be denounced as "positivists" (see Adorno et al. 1969); rather, they combine their empirical linguistics with theory and even theoretical speculation. However, their theoretical speculations are empirically based. After the presentation of the 20 semantic domains and the variable rules of CP production in section 3.3.4.2, I emphasized that my analyses proved that the semantic domains are not static, but rather must be regarded as dynamic and interacting. I pointed out that these dynamic domains can be understood as "program clusters, procedures, scripts" or "functional pathways" (see Pribram 1987:7-12) that speakers use in their speech production. I also noted that combining the individual tables, which attempt to describe the dynamic processes observed within the semantic domains, into one comprehensive table results in a strange kind of "drawing" that looks like a mycelium, or a network to use a different simile. I am aware that the term "network" has been used in the fields of semantics and psycholinguistics (see Collins and Quillian 1969; Lakoff 1987: 116; Miller and Johnson-Laird 1976:272-274; Fox 1975:112, 115-119; Koch 1986:23; Wallace 1989) and that it was also used by Hundius and Kolver (1983:192; see also Unterbeck 1990b:68). I will nevertheless use this network simile in the discussion that follows because I am convinced that it is the most appropriate term to describe the facts observed (see also G. Senft 1991c). On the basis of the analyses presented here, I understand the semantic domains constituted by the CPs types as a network in which the CPs are realized in at least two different ways: First, some are realized within only one semantic domain; I characterize these CPs as elements that are uniquely represented and uniquely localized within the semantic network. Second, some CPs are realized within more than one semantic domain; I characterize these CPs as elements that are multiply represented and multiply localized within the semantic network. These multiply represented and localized CPs can be understood as the network-linking elements, that
326
Classificatory Particles in Kilivila
is, the network ramifications or the network switches that open up and offer the speakers new ways for creative and innovative use of the CPs. Moreover, I assume, on the basis of the results with respect to the actual production of CPs and the processes of language change in progress, that CPs that are uniquely represented and localized elements of the network can change their status and become multiply represented and localized elements. This change of status may be temporary, for example, if a speaker uses this device to pursue certain aims, such as using a certain CP strategically to produce a new metaphor. Alternatively, this change of status may become permament if the speech community accepts the CP as a member of one or more semantic domains that are different from the domain this CP originally co-constituted. This process of status change of an element within the network may also occur in the opposite direction, that is, a multiply represented and localized CP may become a uniquely represented and localized CP if the speech community no longer accepts or uses this CP as a network-linking element. Thus, the dynamics of the network offer the speaker an excellent point of departure in the comprehensive framework of the Sprachspiel (Wittgenstein 1958a 1958b). These dynamics of the network explain the semantic power inherent in the CP system, and they allow consideration of the semantic network established by the CPs as an infinite system—at least in principle. The semantic network which is itself constituted by the semantic domains that in their turn are constituted by the CPs can be described in at least three different ways. First, we could present the network in a linear order. All semantic domains are considered to have the same status and quality within this network, that is, there is no evaluation of the semantic domains that constitute the network. This idea is one-dimensional; however, it has the advantage of offering a model of description with a minimum of basic axioms: we postulate only that a number of CPs establish a number of semantic domains, which, in turn, establish a semantic network. Second, we could present the network in a linear and one-dimensional, but hierarchical, order. Here, we assume that the semantic domains are differentiated with respect to quality or status within the network. Third, we could present the network in a multidimensional hierarchical order. In this case, we would assume that certain semantic domains are located in different levels within the comprehensive hierarchically structured network. This idea, which I can only sketch briefly here, may result in a two-, three-, or even multidimensional model of description. Following the basic idea of the "variety grammar" developed by Klein (1974) and Klein and Dittmar (1979) (see also G. Senft 1982), the dimension of the network in this model depends on intralinguistic and extralinguistic variables (e.g., speaker's gender, age and status, the speech situation) chosen to define the "hierarchy space" the linguist wants to use. With such a multidimensional hierarchy space, the linguist also defines the degree of the netting complexity, the Vernetzungsgrad, to use Vollmer's expression (1988a:136; 1988b:265-267), of the network. That this model is much more complex and needs many
Using Network Models to Describe Classifier Systems
327
more processes of abstraction is evident. I do not want to develop this idea further, but I want to note that such a complex network may well be a good starting point for the linguist's attempt to arrive at a description of language production processes that simulate the actual decision processes and strategies a speaker follows in producing a certain CP. In this monograph, I have extensively developed the first of the three models just presented to describe complex CP systems; I also briefly indicated what the second model might be like. To describe the Kilivila CP system in the framework indicated by the third model would require a separate study; however, such a study would have to be based on insights such as those gained by the research presented here. Of course, these models for describing complex classifier systems can be applied not only to the Kilivila system, but in principle to all classifier languages with a complex inventory of CPs. My work on the Kilivila CP system, especially my idea of presenting the system as a network, has led me to indulge in some general theoretical speculations, and I would like to end this book with these speculations. I have shown that Kilivila native speakers acquire the system of CPs during their maturation process. However, not only the speaker matures; a maturation process also occurs with respect to the CP system: the more CPs speakers acquire and handle in their everyday speech production, the more processes of specification and elaboration of their use of the CP system are observed. These maturation processes seem to be organic. To carry this comparison further, we may characterize the network constituted by the system of CPs as something like an organ with a specific function. We then assume that not only classifier systems but also other systems of speech, as they are described by phonetics, phonology, morphology, syntax, semantics, and pragmatics (as well as some of the speaker's mental abilities and world knowledge), can be regarded as networks, and these networks, in turn, are like organs fulfilling different functions. These networks are interconnected and interact, in hierarchical or in parallel "wiring", as parts of a whole. In our case, this "whole" is language. That this "whole is more than the sum of its parts, and its attributes as a whole are more complex than the attributes of its parts" (Koestler 1978:25; see also Lakoff 1987:273-274) may seem trivial, but is a true insight. I am convinced that this is the reason that linguists are only able to provide descriptions that are complete for the various parts, and that they will never be able to claim that they have described a natural language as a whole. There is always something more that removes itself from the scientist's reach, and this "something more" is what happens in the interaction processes of the interconnected "parts" of the "whole". Wilhelm von Humboldt (1836:CCVII) attempts to compare these processes with the process of crystallization: Die Sprache entsteht, wenn man sich ein Gleichnis erlauben darf, wie in der physischen Natur ein Krystall an den anderen anschieBt, Die Bildung geschieht allmalig, aber nach einem Gesetz, , . . Wenn diese Krystallisation geendigt ist, steht die
328
Classificatory Particles in Kilivila Sprache gleichsam fertig da. Das Werkzeug ist vorhanden, und es falIt nun dem Geiste anheim, es zu gebrauchen und sich hineinzubauen.2
In biologically based philosophy, we find an even better expression than Humboldt's "crystallization" simile. Konrad Lorenz (1973:47-49), in his book Die Ruckseite des Spiegels, (re)introduced the termfulguratio in his biologically founded epistemology:
Theistische Philosophen und Mystiker des Mittelalters haben fur den Akt der Neuschopfung den Ausdruck "Fulguratio", Blitzstrahl, gepragt.... Durch einen etymologischen Zufall, wenn nicht aufgrund tieferer unvermuteter Zusammenhange, trifft dieser Terminus den Vorgang des In-Existenz-Tretens von etwas vorher nicht Dagewesenem viel besser als all die vorerwahnten Ausdrucke. Der Donnerkeil des Zeus ist fur uns Naturforscher ein elektrischer Funke wie jeder andere, und wenn wir an einer unerwarteten Stelle eines Systems einen Funken aufblitzen sehen, so ist das erste, woran wir denken, ein KurzschluB, eine neue Verbindung. Wenn z.B. zwei voneinander unabhdngige Systeme zusammengeschaltet werden ... so entstehen damit schlagartig vollig neue Systemelgenschaften, die vorher nicht, und zwar auch nicht in Andeutungen, vorhanden gewesen waren. Genau dies ist der tiefe Wahrheitsgehalt des mystisch klingenden, aber durchaus richtigen Satzes der Gestaltpsychologen: Das Ganze ist mehr als seine Teile.3 Let me summarize the train of thought presented here: I assume that language is the whole that "fulgurates" in the process of interconnecting the different and more or less distinct parts that the linguist (the sociolinguist, the ethnolinguist, the psycholinguist, the typologist from her or his point of view) describes on the levels of phonetics, phonology, morphology, syntax, 2 "Language emerges, if the simile is permitted, like one crystal buds off from another in physical nature. The formation occurs gradually; however, it is governed by a rule.... If this crystallization process ends, the language stands there, practically ready. The tool is available, now it falls to the mind to use it and to build itself into it" (Humboldt 1836: CCVII [my translation, G.S.]). 3 "Theistic philosophers and mystics of the Middle Ages coined the term fulguratio, 'flash of lightning,' to denote the act of creation ...." By an etymological accident or perhaps through deeper, unsuspected associations, this term is far more appropriate than those mentioned above for designating the coming into existence of something previously not there. A thunderbolt from Zeus is for the scientist an electric spark like any other, and if we see a spark at an unexpected point in a system, the first thing we think of is a short circuit, a new connection. "If, for example, two independent systems are coupled together ... entirely new, unexpected system characteristics will emerge, of whose appearance there was previously not the slightest suggestion. This is the profound truth behind the Gestalt psychologists' principle, mystical in tone but absolutely correct, that the whole is more than the sum of its parts." (Lorenz 1977:29-30 [translated by Ronald Taylor].)
Using Network Models to Describe Classifier Systems
329
semantics, and pragmatics. These parts clearly constitute what can be understood as complex networks. We can achieve a more or less complete description of the network parts; however, we cannot achieve a description of the result of the interconnection of these parts—the "whole" of language—as long as we are ignorant of what happens in the process of fulguration. I hope that I have reached my goal, namely, to describe as comprehensively as possible one part of the language I have been working with for almost seven years: the network system of classificatory particles within Kilivila—the "CP-part" of the "Kilivila-whole."
Appendix
A
Consultants in 1982/1983 Study Elicitation Test Consultant Data Consultant no. and approximate age
Name
Clan
Group I Females
1 (6 y)
2 (4 y) 3 (4 y) 4 (7 y) 5 (5 y) 6 (6 y) Males 7 (4 y) 8 (5 y) 9 (4 y) 10 (6 y) 11 (6 y) 12 (7 y)
Igiobibila Vesali Omnava Payaya Kwelubituma Inukwala
Malasi Lukulabuta Lukuba Malasi Malasi Malasi
Topsikauya Yabilosi Yogima Towesei Milavatu Morona
Lukulabuta Lukwasisiga Lukwasisiga Lukuba Malasi Lukulabuta
Average age: females = 5.3 y, males = 5.3 y, All = 5.3 y. Group II Females 13 (11 y) 14 (8 y) 15 (11 y) 16 (11 y) 17 (12 y) 18 (14 y) Males 19 (13 y) 20 (9 y) 330
Bolubatau Olopola Imkubula Emi-Iborogu Namyogai Iluboku
Lukwasisiga Lukwasisiga Malasi Malasi Malasi Lukwasisiga
Pulia Tosulebu
Lukwasisiga Malasi
Consultant: 1982/1983 Study
331
Table (continued) Consultant no. and approximate age
Name
Clan
21 22 23 24
Buligesi Gumsakapu Morona Luluwasikweguyau
Lukwasisiga Lukwasisiga Malasi Malasi
(12 y) (14 y) (10 y) (8 y)
Average age: females = 11.2 y, males = 11.0 y, All = 11.1 y. Group III Females 25 (16 y) 26 (17 y) 27 (20 y) 28 (18 y) 29 (16 y) 30 (19 y) Males 31 (19 y) 32 (20 y) 33 (16 y) 34 (15 y) 35 (16 y) 36 (15 y)
Namnabai Itakeda Kapatu Asinata Bomsamesa Ibonoma
Malasi Malasi Lukwasisiga Lukwasisiga Malasi Malasi
Gayoboda Moagawa Kaluvalu Pwaraesa Kwelava Moromata
Malasi Lukwasisiga Lukwasisiga Malasi Lukwasisiga Lukulabuta
Average age: females = 17.7 y, males = 16.8 y, All = 17.3 y. Group IV Females 37 (26 y) 38 (27 y) 39 (29 y) 40 (28 y) 41 (32 y) 42 (28 y)
Naukwatai Yebwaku Bwetagava Dakevau Bokamata Igogosa
Lukwasisiga Malasi Malasi Lukulabuta Malasi Lukwasisiga
Males 43 (30 y) 44 (35 y) 45 (30 y) 46 (26 y) 47 (25 y) 48 (27 y)
Toybokwatauya Tokuyumila Bwetadou Tokwakuva Mogega Moagwana
Lukwasisiga Malasi Lukulabuta Lukwasisiga Malasi Lukwasisiga
Average age: females = 28.3 y, males = 28.8 y, All = 28.6 y. Group V Females 49 (56 y) 50 (44 y) 51 (48 y) 52 (74 y) 53 (58 y) 54 (39 y)
Vadomna Sogeya Kadawaya Sedaka Sibwesa Rita-Isadoga
Malasi Lukwasisiga Malasi Malasi Lukwasisiga Lukwasisiga
Classificatory Particles in Kilivila
332
Table (continued) Consultant no. and approximate age
Name
Clan
Males 55 (62 56 (37 57 (45 58 (58 59 (75 60 (48
Weyei John-Bomyoyewo Nusai Kasilasila Vaka'ila Vapalaguyau
Malasi Malasi Malasi Lukwasisiga Malasi Malasi
y) y) y) y) y) y)
Average age: females = 53.2 y, males = 54.2 y, All = 53.7 y.
Overall Corpus of Kilivila Speech Data Sex M M M M M M M M M M M M M F F F F M M M F M M M M M M M M M M
Age group I I I II II II II III IV IV IV
No./no. in elicitation test 1/8
2/10
3 1/19 2/20
3 4 1 1/43 2/45
Age
Name
Clan
(5 y) (6 y) (7 y) (13 y) (9 y) (9 y) (12 y)
Yabilosi Towesei Dudauvelu Pulia Tosulebu Dauya Mosuelebu Stanley-Uveaka Toybokwatauya Bwetadou Kalavatu Kalitaiga Tolivalu Vadomna Sogeya Kadawaya Sibwesa Weyei John-Bomyoyewo Vapalaguyau Bomesa Bwema'utila Gerubara Kilagola Katubai Mokeilobu Mokopei Tokonupei Tomalala Tosulala Yaurabina
Lukwasisiga Lukuba Malasi Lukwasisiga Malasi Malasi Malasi Lukwasisiga Lukwasisiga Lukulabuta Lukwasisiga Malasi Malasi Malasi Lukwasisiga Malasi Lukwasisiga Malasi Malasi Malasi Malasi Lukulabuta Lukwasisiga Malasi Malasi Malasi Lukwasisiga Malasi Lukulabuta Malasi Malasi
(20 y) (30 y) (30 y)
IV
3 4 5
(30 y) (28 y) (28 y)
1/49 2/50
(56 y)
V
3/51
(48 y)
4/53
(58 (62 (37 (48 (60
IV V
V V V V
V
V V
V
V
V
V V
V V V V
5/55 6/56 7/60
8 9
10 11 12 13 14 15 16 17 18
(44 y) y) y) y) y) y)
(57 y) (39 y) (63 y)
(45 y) (43 y) (60 y)
(56 y) (39 y) (41 y) (65 y)
Appendix
B
Number of CP Tokens Produced for Each CP Type by Text and Word Class for Consultants in Corpus of Kilivila Speech Data Age group, consultant no., name
CP tokens Text category
CP type
I 1/8 Yabilosi
Fairy tale
kwe
I 2/10 Towesei
Fairy tale
bukwa kai ke kwe na pila pwanina to/te ya dumia ke na pila to/te
Nursery rhymes
Interview I 3 Dudauvelu
Fairy tale
kada ke kova kwe na to/te ya
N
D
A
4
2
IP
1 l 3 1 4 2
3 1
2 2 2 1
10 6
20
1 13 3
1 4 3 1
2
2 2
5 5 7
1
2 3 35 9 7
6 1
— numeral; D — demonstrative pronoun; A — adjective; IP = interrogative pronoun.
333
334
Classificatory Particles in Kilivila
Table (continued) CP tokens
Age group, consultant no., name
Text category
CP type
II 1/19 Pulia
Description
ke kwe tolte ya
II 2/20 Tosulebu
II 3 Dauya
Fairy tale
Fairy tale
Interview II 4 Mosuelebu III 1 Stanley-Uveaka
Interview Public speech
Interview
IV 1/43 Toybokwatauya
Fairy tale
Myth
ke kwe na tolte
kabulu ke kwe na pila tolte kwe tolte tolte ke kwe palpila tolte kwe na tolte ke kwe na tolte ke kwe palpila tolte uva
Magic Nursery rhymes
Nursery rhymes Songs Description
N 1 1 1
tarn
ke kwe
1
A
IP
1 2 1
1 2
1 3
4
4 1
5 1
3 1 4 1
8 22 1 4
1
7 1
8
13 2
1 1
2 4 4
4
1 1
10 3
2
3 8
2
1
2
3 16 4
6 3
2 1 3
1
1
ya ke
yule ke kwe palpila tolte ke ke palpila kasa
D
1
1 2 1 4 1 1 6 2
3 1
1 1 2 2
1 1
5 15 4
6
1
Number of CP Tokens Produced for Each CP Type
335
Table (continued) Age group, consultant no., name
IV 2/45 Bwetadou
IV 3 Kalavatu
CP
Text category
Description
Nursery rhymes Description
Description
CP type na to/te kai ke kwe liku ke to/te kauya ke kwe lipu palpila tolte ya ke pila ta tolte uva ya yuma
IV 4 Kalitaiga
Description
ke kwe
IV 5 Tolivalu
Public speech
gula
V 1/49 Vadomna
Description
V 2/50 Sogeya
Description
ke kwe ke kudu kwe tolte buda kwe tolte kwe na bubwa ke kova kwe na tolte ya
V 3/51 Kadawaya
Public speech Description
V 4/53 Sibwesa
V 5/55 Weyei
Interview
Public speech Magic
kwe na tolte kwe
N 1 1 3
tokens D
A
IP
4 1 5
1 4 1
4 1 1 4
2 1 5 1
1
1
1 1 2
2 1 2
1
1
3
1
7 1 1 1 1 1
1
1 2
1
2 4 10
1 1 1 1
1
3 2
1 2 1 5 20 5 1 2 1 1 1
1 3 2
14
336
Classificatory Particles in Kilivila
Table (continued) CP tokens
Age group, consultant no., name
Text category
CP type
V 6/56 John Bomyoyeva
Description
ke kwe
V 7/60 Vapalaguyau
Interview
gula ke kwe kwela na tolte
V 8 Bomesa
Songs
V 9 Bwema'utila
Public speech
V 10 Ata-Gerubara
Public speech Myth Prayer
V 11 Kilagola
Public speech
Public speech
Myth
Joke
Magic Description Description
ke na ke
N 3 2
19 1 1
2 1 1
kwe na tolte ya buda gula kada ke kwe na tolte buda ke kudu kwe kweya na pila tolte ya ke na palpila tolte ya kwe ke
5
na
A
1 2 1 1 3 1 1
3
IP
1 1 4
2 1
ke kwe tolte na tolte tolte
kwe ke kwe
D
1 1 2 1
3
4
5
9 2 1
2 2 1 1 1 5 1 1 1 8 1 11 3 2 1
5
20 1 10
8
30
1
4
3 1
1 2 14 3 1
5 1 1 1
12 2 25
2
1
2
3 1 7
9 3 8 2 3
2
4 1 1 2
9 1
1
Number of CP Tokens Produced for Each CP Type
337
Table (continued) Age group, consultant no., name
V 12 Katubai V 13 Mokeilobu
CP tokens Text category
Public speech Public speech
Description
V 14 Mokopei
V 15 Tokunupei
Myth
Description Description Interview
V 16 Tomalala
Public speech Public speech
Description V 17 Tosulala
Description
CP type palpila tolte utu kwe ke kwe na tolte ke kwe tolte ya gudi kali ke kwe kweya na palpila sisi tolte
kauya kwe tolte ke kwe ta buda gudi ke kwe na tolte kwe kabulo kadalkeda ke kwe kwela na palpila tolte ya ka kwe bililo bwa kwabila
N
1 1
D
A
2 4
7
2
4
3
4 1
6 3 1 1 1 1 2 1
1
1 11 1 1
2 6
2 1
1 19 15
6
8 2
11
1 1
7 2
IP
1
2 3 2 1 1 1 2 2
21
1 1 2 10 1
3 3
1 13 36 14
1 3
7
4
2
1 1
5 10
2
4 9 3
8 39 1 2 1
4 2 1
3 2
2
Classificatory Particles in Kilivila
338
Table (continued) Age group, consultant no., name
CP tokens Text category
Description
V 18 Yaurabina
Public speech Public speech
CP type kai ke kabulu kwe liku to/te uva ya bubwa kasa ke kova kwe to/te yeni kwe to/te utu kwe to/te
N
4 1
2
4
D
A
6 35
2
2 2 6
1 1
5
8 3 1
3 2
1 2 1 2 1
5
3 3 1 4
3 2
4 3
1 1
9 1
1
1 4
2 2 5
CPs produced but not assignable to individual consultants Nursery rhymes
Songs
ke kwe to/te bubwa kasa na sam ya 0 buda ke kwe na pila to/te
6
5 1 2 1 3 2
kwe
na to/te
1
1 1 6
1 1
7 2
ya
Interview Background
2
7
4 3
3 1
Public speech
kwe
na to/te
3 1 1
Prayer
ya buda
1
kai
kwe to/te
2 1
1 1
2 1 1 13 1 8
IP
Appendix
C
Consultants: 1989 Restudy
Restudy Question: Do Kilivila native speakers accept the weighting and ordering of the semantic domains as expressed by the hierarchy index in Table 3.4.22?
Age group (years)
Consultant name
II (8-14)
Igiobibila Ebutu Kalibokio Dudauwelu Mokeimeku Moyadoga
III (15-20)
IV (21-35)
V (36-75)
Senubesa Payaya Namyogai Gumsakapu Morakum Dauya Kapudokoya Dakevau Yebwaku Kalitaiga Tolivalu Tova'ula Igiova Nakivila Inadila Motaesa Vapalaguyau Vasopi
Sex f f f m
m
Age (years) 12
9 10 13 11
m
10
f f f m m m f f f m m m f f f m m m
20 15
18
20 17 15 26 34 33 34
34 30
70 36 48 62 54 58
Clan Malasi Lukwasisiga Malasi Malasi Malasi Malasi Lukuba Malasi Malasi Lukwasisiga Malasi Malasi Malasi Lukulabuta Malasi Malasi Malasi Malasi Malasi Lukulabuta Malasi Malasi Malasi Malasi
339
Classificatory Particles in Kilivila
340
Restudy Question: Do Kilivila native speakers accept the semantic domains constituted by the CP types proposed in section 3.3.4.2?
Age group (years)
Consultant name
I(4-7)
Bokarawana Nakilim Vadomna Mobiliuya Yaurabina Mopili
II (8-14)
III (15-20)
Kwelubituma Igiobibila Olopola Towesei Tonakola Tomdoya Iluboku Namyogai
Emi
Kwelava Dukuta'isi Tom IV (21-35)
V (36-75)
Bomsamesa Itakeda Kapudokoya Gayoboda Moagawa Gumadaka Sogeya Vadomna Kadawaya Kilagola Bwetadou Mwasei
No.
Sex
— — — — — — 1
f f f
2
3
4 5 6
7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24
Age (years) 4 6
Clan Lukulabuta Lukulabuta Malasi Malasi Malasi Lukwasisiga Malasi Malasi Lukwasisiga Lukuba Malasi Lukulabuta
m m m f f f
4 4 4 5 11 12 14
m
12
m m f f f
20
Lukwasisiga Malasi Malasi Lukwasisiga Lukulabuta Lukwasisiga
22
Malasi Malasi Malasi Malasi Lukwasisiga Lukuba Lukwasisiga Malasi Malasi Malasi Lukulabuta Malasi
m m m f f f
m m m f f f m m m
10 11 18 17 20 15 18 23 26 25 26 23 50 62 54 69 36 51
Consultants: 1989 Restudy
341
Restudy Question: Do Kilivila native speakers accept, reject, or modify formulated in section 3.3.4.2?
Age group (years)
Consultant name
II (8-14)
Igiobibila Bomtula Bomlisi Moyadoga Toyogima Baigaega Imkubula Iluboku Ilitula Mwelabusi Tosuelebu Dauya
III (15-20)
IV (21-35)
V (36-75)
Bwetagava Ibonoma Bokarawana Moagawa Tova'ula Kalitaiga Inoma Inadila Kadawaya Gerubara Tokunupei Kilagola
No. 1 2
3 4 5 6 7
8 9
10 11 12
13 14
15 16 17 18 19 20 21
22 23 24
Sex
Age (years)
f f f
12 10 13
m m f f f
12
m
m m m f f f m m m f f f m m m
10
14 17 20 19 16 15 15 35 25 30 26 30 34 39 48 54 45 62 69
the variable rules
Clan Malasi Malasi Malasi Malasi Lukwasisiga Lukwasisiga Malasi Lukwasisiga Malasi Malasi Malasi Malasi Malasi Malasi Malasi Lukwasisiga Malasi Malasi Lukwasisiga Malasi Malasi Lukwasisiga Malasi Malasi
Restudy Question: Do Kilivila native speakers accept, reject, or modify the variable rules formulated in section 3.3.4.2?
Domain la
Concept Persons Children
Corpse Spirit 1b
Limb
Hand
Tooth
Phrases presented in language evaluation test O davalusi esisusi tomanabweta tauwau namanabweta vivila gudimanabweta gugwadi O davalusi esisusi tomanabweta tauwau tomanabweta vivila tomanabweta gugwadi. minana tomata minana kosi
Gloss In our village live CP-beautiful men, CP-beautiful women, (and) CP-beautiful children.
this-CP-this corpse this-CP-this spirit
makweyana imitabogu makwena imitabogu makena imitabogu mayamna yamala makweyana yamala makwena yamala mtona yamala mayana yamala maka'ina kudula makuduna kudula makwena kudula makena kudula
this-CP-this my finger
this-CP-this her/his hand
this-CP-this her/his tooth
2
Animal Shell
minana mauna makwena vigoda *minana vigoda
this-CP-this animal this-CP-this shell
3a
Group/team
makwena boda mabudana boda madelina deli makwena deli mtona deli minana deli makasana deli
this-CP-this group
Group move
this-CP-this group on the move
Shoal
Batch/fish
Fish/strings
3b
Heap
Plantation
Fruit cluster
Cluster/bundle
Bundle/wrapped
mabudana boda bodavakaveaka makwena boda bodavakaveaka mayuvana boda bodavakaveaka makena boda bodavakaveaka makasana boda bodavakaveaka masuyana yena budubadu esisusi minasina yena budubadu esisusi mawelana yena budubadu esisusi minasina yena isuisi mawelana yena isuisi maoylana yena isuisi makwena gugula magulana gugula matamna gugula mabudana gugula makwena po'ula mapo'ulana po'ula magulana po'ula makena kavailuva mayana kavailuva mabukwana kavailuva makwena kavailuva masana kavailuva mapo'ulana kavailuva maluvana kavailuva madulina duli makwena duli makena duli malubana duli elubisi makapwana duli elubisi makwena duli elubisi
this-CP-this big shoal
this-CP-this batch of fish-many are there
this/these-CP-this/these fish they are on strings
this-CP-this heap
this-CP-this plantation
this-CP-this (cluster of) fruit
this-CP-this cluster/bundle
this-CP-this bundle they wrapped (it) up
In general, phrases are presented here with demonstrative pronouns only. I present these phrases to give the reader a better understanding of the language evaluation test. For the glosses of these phrases CPs are not translated.
Table (continued) Domain
Concept Small bit
Bit/piece
Part/piece
Scrap
Cluster/bananas Conical bundle
Phrases presented in language evaluation test ma'utuna tobaki pikekita masina tobaki pikekita mavilina tobaki pikekita makwena tobaki pikekita mabwana tobaki pikekita makena tobaki pikekita magumna gum makwena gum makena gum mayana gum makasana gum ma'utuna gum mapilana sitana mayana sitana makena sitana makwena sitana mavilina sitana ma'utuna sitana ma'utuna sitana ebwabusi makena sitana ebwabusi mabwana sitana ebwabusi mabubwana sitana ebwabusi makena usi duli galayomala makilana usi duli galayomala makwena usi duli galayomala mammwana uli duli bwena makena uli duli bwena magubana uli duli bwena
Gloss this-CP-this little (bit of) tobacco
this-CP-this small piece
this-CP-this part
this-CP-this part (which) they cut off
these-CP-these bananas many bundles
these-CP-these taro (a) good bundle
Nut bunch
Sheaf
Bundle of four things
Handful
makwena uli duli bwena maluvana uli duli bwena mabudana uli duli bwena masana buva budubadu makena buva budubadu makwena buva budubadu maluvana buva budubadu mapo'ulana buva budubadu masipuna sipu makwena sipu mayana sipu makena sipu makwena vavagi makala kwevasi uk makena vavagi makala kevasi uk mayulaina vavagi makala yulaivasi uk mayenina pwaka
these-CP-these betelnuts many (of them)
this-CP-this sheaf
this-CP-this thing (looking) like CP-four hooks
this-CP-this lime
4
Anything
makwena vavagi avaka avaka makena vavagi avaka avaka
this-CP-this thing whatever (it may be)
5
Span
ma'uvana kuvi makena kuvi makwena kuvi matamna kuvi mayumana pepe'u
this-CP-this "kuvi" (-type of) yams
6
Night
makwena bogi mabogina bogi makwena yam makalana yam matutana yam mayamna yam makwena tuta matutana tuta
this-CP-this night
Length
Day
Time
this-CP-this small basket
this-CP-this day
this-CP-this time
Table (continued) Domain
7
Concept
Phrases presented in language evaluation test
Day/part
masivana tuta ma'utuna tuta matutana tuta mapilana tuta mameilana yam
Village sector
Place/area Sea portion
Land plot
Corner/garden
Mountain
makabulona kabuluyuvela makena kabuluyuvela makwena kabuluyuvela ma'utuna kabuluyuvela makwena valu mavilona valu makwena bwalita (mtona wala magila epola) mabwalitana bwalita (. . .) ma'utuna bwalita (. . .) masiwana bwalita (. . .) makubilana kwabila makwena kwabila makena kwabila mapilana kwabila manununa nunula bagula makwena nunula bagula makabulona nunula bagula ma'utuna nunula bagula makoyana koya makwena koya
Gloss
this-CP-this part of the day this-CP-this village sector
this-CP-this place this-CP-this part of the sea (where this man only has the right to fish)
this-CP-this land
this-CP-this corner of the garden
this-CP-this mountain
8
Swamp Moon Star Area/authority River/creek
madumiana dumia minana tukukona minana manova malikuna kabuluyuvela guyau esisu mavasina yeyelu, vaya, bwalita
this-CP-this this-CP-this this-CP-this this-CP-this this-CP-this
Cut
magiwina giwi makena giwi makwena giwi makasana giwi maponana pwanina makwena pwanina manigwana pwanina mabubwana sitana avaka ebwabusi makena sitana avaka ebwabusi mabwana sitana avaka ebwabusi ma'utuna sitana avaka ebwabusi makwena sitana avaka ebwabusi mabubuna bobu ebwabusi makwena bobu ebwabusi makena bobu ebwabusi mabwana bobu ebwabusi mayana bobu ebwabusi mabubwana bobu ebwabusi miyana yoyu masisina yoyu makena yoyu masina yoyu mapilana yoyu mavilina tobaki dualilia makena tobaki dualilia ma'utuna tobaki dualilia makwena tobaki dualilia mabwana tobaki dualilia
this-CP-this cut
Punctured
Cut off
Cut across
Flexible/thin
Untwisted
swamp moon big star village sector (where) the chief lives small creek, river, sea
this-CP-this hole
this-CP-this part that they cut off
this-CP-this log they cut off
this-CP-this palm branch
this-CP-this (bit of) straight (stick of) tobacco
Table (continued) Domain
9
Concept Protuberances
Row/line
Row
Kneaded/dot/drop
10
Wood Trees
Bough/cut off Bough/branch Wooden dishes
Phrases presented in language evaluation tesf makabulona matala peni makena matala pent ma'utuna matala peni mabwana matala peni makasana kasa makwena kasa makena kasa magilina vakala mayana vakala makwena vakala mavilina vakala makasana vakala mavakalana vakala manotuna notu makwena notu makena kai mabwana kai makesina nuya, buva, seda, veva mabwasina nuya, buva, seda, veva masisina sisila miyana sisila makena keyala deli kelima kotolu malilana keyala deli kelima ketolu makena kaboma maluvana kaboma makwena kaboma
Gloss this-CP-this tip of the pencil
this-CP-this line this-CP-this belt of spondylus shell discs (in a row)
this-CP-this dot this-CP-this (piece of) wood these-CP-these palm-, betelnut-, nut-, and mangotrees this-CP-this (cut off) branches this-CP-this spear with CP-five CP-four (= nine) heads this-CP-this wooden dish
11
Carving/human Tool
String
Basket Pot/potlike Fish trap 12
Yams/basketful
13
Compartment/foodhouse
Compartment/canoe/foodhouse Compartment/creel
minana tokwalu odabala lagim
this-CP-this carving on top of the canoe board
makavina kavi makena kavi makwena kavi mapilana kavi makena wela makwena wela mapilana wela maoylana wela matanalmapetana peta makwena peta mayana peta makwelana kwena makwena kwena makena kwena makauyana kivaya makwena tetu tala tetu matamna tetu matetuna tetu
this-CP-this tool
makabisina liku makena liku makwena liku mabwana liku ma'utuna liku mapilana liku malikusina waga kena liku makwesina waga kena liku makesina waga kena liku malipuna kivaya
this-CP-this foodhouse
this-CP-this string of fish
this-CP-this basket
this-CP-this pot this-CP-this fish trap this-CP-this yams
this-CP-this (part of a) canoe or foodhouse this-CP-this (part of a) fish trap
Table (continued) Phrases presented in language evaluation test
Domain
Concept
14
Door/entrance
maduyana taboda mavayana taboda makadana taboda mayana taboda makwena taboda makena taboda
this-CP-this door
15
Fire/oven
makovana kova makena kova makwena kova makumlana kumkumla makwena kumkumla makadana keda malilona lola makwena lola mayana lola makadana lola makena lola mabililona lola makalina kaikela baolaola
this-CP-this fire
Earth oven
16
Road/track Walk/journey/trip
Paddle strike
17
Part/song
maninana wosi mapilana wosi makwena wosi mamegwana wosi ma'utuna wosi makasana wosi
Gloss
this-CP-this earth oven this-CP-this road this-CP-this journey
(with) this-CP-this paddle I will paddle this-CP-this song
Part/song/magic
Division/magic
18
Stone blade
Ginger
19
Band/fibers
Belt/shells 20
Name
mameilasina wosi deli megwa mapilasina wosi deli megwa mavosisina wosi deli megwa makesina wosi deli megwa makasasina wosi deli megwa makwesina wosi deli megwa ma'utusina wosi deli megwa masisina megwa
these-CP-these songs and magical formulae
makaina beku mabekuna beku makwena beku masamna neya
this-CP-this stone blade
this-CP-this (piece of) magic
this-CP-this ginger
makwena doba makuduna doba madobana doba mavakalana vakala
this-CP-this grass-skirt
ma'igana yegila makwena yegila
this-CP-this name
this-CP-this belt
Appendix
D
Some Speculations on the Origin of Classifers Most linguists who study the phenomenon of classifiers cannot refrain from speculating on the origin of these formatives. Adams (1989:193-194) gives an excellent general account of the most interesting hypotheses proposed to date (see also Asmah 1972; de Leon 1988:128, 134-136, 141-142, 168-170; Seiler 1983; Unterbeck 1990a:8; 1990b:86, 90). On the basis of my research on Kilivila, I speculate in this appendix on the origin of the Kilivila classifiers. I assume that probably most, if not all, Kilivila CPs are traceable to nouns (see G. Senft 1993). This is a rather obvious hypothesis, considering that quite a few CPs are "repeaters". A repeater is a form that can function as a CP as well as a noun, or one might say, a repeater is a noun that functions as its own classifier. Repeaters are found in many classifier languages,1 and the idea that CPs have evolved from nominals is far from original.2 Nevertheless, let me develop this hypothesis briefly. If we look again at Table 3.3.1 in section 3.3.1 and compare the CPs there with the lexical entries given in G. Senft (1986:185-430), we find the following: The CPs or CP variants beku, bogi, boda, bwalita, doba, duli, dumia, gula, yegila, keda, kasa, kova, kwena, koya, megwa, nigwa, peta, po'ula, siva, tetu, tuta, vaya, vakala, vosi, and yam are repeaters. These 25 repeaters represent 28% of the 88 CPs discussed in detail in this book. The CP gudi is a shortened form of the noun gwadi; the CP variant iga is a shortened form of the noun yegila (note that we also documented the CP variant yegila); the CP ke is a shortened form of the noun kai (note that in highly formal situations the CP variant kai is produced); the CP kumla is a shortened form of the noun kumkumla; the CP variants kwoila/kwela/kway/ kwaila/kwaikwa are modified forms of the noun kwena (note that we also 1 See, for example, Asmah (1972:88-90); Barz and Diller (1985:168, 174); and the references in section 1.2 of the present volume. 2 See also Carpenter (1986:18; 1992:139) and Lee (1987:404). Of course, classifiers may also originate in other word classes. Seiler (1986:17) argues that in Imonda, a Nonaustronesian language of the Trans-New Guinea phylum," erstwhile full verbs were semantically depleted and reanalysed as classifiers". See also Seiler 1983. The influence of classifier systems on other classifier systems and the consequences for CP systems of borrowing from other languages are discussed by Adams (1991).
352
Some Speculations on the Origin of Classifiers
353
documented the CP variant kwena); the CP na in its meaning 'person of female gender' is probably a shortened and modified form of the noun vivila (in the biga galanani variety of Kilivila, the noun is realized as vivina); the CP variants pwanina/pona/ponina/ponu/polu/pwana are shortened or modified forms of the noun pwanana; the CP te/to is a shortened form of the noun tau (note that in highly formal situations the CP variant tau is produced); and the CP vilo is a modified form of the noun valu. These nine cases cover seven more CP types that can be traced back to nouns; these seven CP types represent another 8% of the 88 CPs described in detail in this book. The CP kabulo/kabulu in its meanings 'protuberances' and 'cape, point' can be traced back to the nouns kabulula 'point, ledge', kabunu 'point', and kabulu-PP IV 'nose'; the CP kauya can be traced back to the noun kauya 'woven basket' (however, the CP has a more specific meaning than the original noun); the CP kudu in its meaning 'tooth' can be traced back to the noun kudu-PP IV; the CP liku can be traced back to the noun liku 'big foodhouse' (however, the CP has a more specific meaning than the original noun); and the CP yuma, yam in its meaning 'hand' can be traced back to the noun yama-PP IV. These five CPs represent another 6% of the 88 CPs described in detail here. In summary, 37 of the 88 CPs described in detail in this book can be readily traced to nouns. Thus, the hypothesis presented here seems to be highly plausible for 42% of the described CPs. Malcolm Ross, in discussing this aspect of my research with me, suggested that I reconstruct the protoforms for the Kilivila CPs to find more evidence for this hypothesis (Ross 1989: personal communication). This is a very interesting proposal. However, I have some further evidence to support my hypothesis. Examination of the tables in section 3.3.4.2, which present the dynamic processes within the semantic domains, shows the following: the CP kweya/kwaya/keya supersedes the CP yam/yuma/yama and the CP kcti supersedes the CP kudu (see Table 3.4.1.2a); the repeater peta is superseded by its shortened variant ta and the repeater yegila is superseded by its shortened variant iga; the CPs bogi, koya, po'ula, siva, yam, and to a certain extent the CP boda, are superseded by the general CP kwe; the CP liku and to a certain extent the CP duli, are superseded by the general CPs ke and kwe; the CP kabulo is superseded by the CP ke; the position of the CP gula is affected by the CPs tam and kwe; and the CPs beku, bwalita, doba, dumia, megwa, nigwa, tetu, tuta, vaya, vakala, vosi, and kauya play only a marginal role within the system of Kilivila CPs. These additional observations for 26 of the 37 CPs that could be traced to noun forms lead me to draw the following conclusion. I assume that most, if not all, Kilivila CPs originate in nouns (for a more general formulation of this idea, see Lee 1988:212, 225, 235). Originally, the CP system contained repeaters only. Over time, these repeaters were changed and modified, usually by processes of phonological reduction. Finally, CPs that can be easily traced to nouns have been superseded by the general CPs or by CPs that have undergone so many changes that their nominal origin is difficult or perhaps impossible to trace.
This page intentionally left blank
References Adams, Karen L. 1986. Numeral classifiers in Austroasiatic. In Noun classes and categorization, ed. Colette Craig, 241-262. Amsterdam: John Benjamins. Adams, Karen L. 1989. Systems of numeral classification in the Mon-Khmer, Nicobarese and Aslian subfamilies of Austroasiatic. Pacific Linguistics, Series B, No. 101. Canberra: Australian National University. Adams, Karen L. 1991. The influence of non-Austroasiatic languages on numeral classification in Austroasiatic. Journal of the American Oriental Society 111(1):62-81. Adams, Karen L. 1992. A comparison of the numeral classification of humans in Mon-Khmer. Mon-Khmer Studies 21:107-129. Adams, Karen L., and Nancy F. Conklin. 1973. Toward a theory of natural classification. In Papers from the Ninth Regional Meeting, Chicago Linguistic Society, eds. Claudia Corum, T. Cedric Smith-Stark, and Ann Weiser, 1-10. Chicago Linguistic Society, University of Chicago, Chicago, I11. Adams, Karen L., and Nancy F. Conklin. 1974. On the numeral classifier in Thai. Paper presented at the 7th International Conference on Sino-Tibetan Language and Linguistic Studies, Atlanta, Ga, October 19, 1974. Adams, Karen L., Alton L. Becker, and Nancy F. Conklin. 1975. Savoring the differences among classifier systems. Paper presented at the 8th International Conference on Sino-Tibetan Languages and Linguistics, Berkeley, Calif., October 1975. Adorno, Theodor W., Ralf Dahrendorf, Harald Pilot, Hans Albert, Jurgen Habermas, and Karl R. Popper, eds. 1969. Der Positivismusstreit in der deutschen Soziologie. Darmstadt, Germany: Hermann Luchterhand. Allan, Keith. 1977. Classifiers. Language 53:285-311. Alungum, John, Robert J. Conrad, and Joshua Lukas. 1978. Some Muhiang grammatical notes. In Workpapers in Papua New Guinea languages, ed. Richard Loving, 89-130. Ukarumpa, Papua New Guinea: Summer Institute of Linguistics. Aristotle. 1976. The Ethics of Aristotle. The Nicomachean Ethics. Translated by J. A. K. Thomson. Harmondsworth, England: Penguin. Asmah, Jaji Omar. 1972. Numeral classifiers in Malay and Iban. Anthropological Linguistics 14(3): 87-96. Austen, Leo. 1939. The seasonal gardening calendar of Kiriwina, Trobriand Islands. Oceania 9:237-253. Austen, Leo. 1949/50. A note on Dr. Leach's "Primitive calendars." Oceania 20:333-335. 355
356
References
Baldwin, Bernard. No date. Biga Boyowa, Catholic Mission Gusaweta, Papua New Guinea. Ms. Barron, Roger. 1982. Das Phanomen klassifikatorischer Verben. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 1: Bereich und Ordnung der Phanomene, eds. Hansjakob Seiler and Christian Lehmann, 133-146. Tubingen, Germany: Narr. Barron, Roger, and Fritz Serzisko. 1982. Noun classifiers in the Siouan languages. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 2: Die Techniken und ihr Zusammenhang in Einzelsprachen, eds. Hansjakob Seiler and Franz-Josef Stachowiak, 85-105. Tubingen, Germany: Narr. Barz, R. K., and A. V. N. Diller. 1985. Classifiers and standardisation: Some South and South-East Asian comparisons. In Papers in South-East Asian linguistics, No. 9, ed. Daniel Bradley, 155-184. Canberra: Australian National University. Bechert, Johannes. 1982. Grammatical gender in Europe: An areal study of a linguistic category. Papiere zur Linguistik 26: 23-34. Bechert, Johannes. 1988. The structure of the noun in European languages. Paper presented at the Workshop on Typology of Languages in Europe. Rome, Italy, January 1988. Becker, Alton L. 1975. A linguistic image of nature: The Burmese numerative classifier system. Linguistics 165: 109-121. Becker, Alton L. 1986. The figure a classifier makes: Describing a particular Burmese classifier. In Noun classes and categorization, ed. Colette Craig, 327-343. Amsterdam: John Benjamins. Bell-Krannhals, Ingrid. 1987. Nahrung aus dem Meer: Fischfang an den Kusten Neuguineas. In Neuguinea: Nutzung und Deutung der Umwelt, ed. Mark Munzel, vol. 1, 239-261. Frankfurt am Main, Germany: Museum fur Volkerkunde. Benton, Richard A. 1968. Numeral and attributive classifiers in Trukese. Oceanic Linguistics 7:104-146. Berlin, Brent. 1968. Tzeltal numeral classifiers: A study in ethnographic semantics. The Hague: Mouton. Berlin, Brent, and A. Kimball Romney. 1964. Descriptive semantics of Tzeltal numeral classifiers. American Anthropologist 66(3):79-98. Berlin, Brent, Dennis E. Breedlove, and Peter H. Raven. 1969. Folk taxonomies and biological classification. In Cognitive Anthropology, ed. Stephen A. Tyler, 60-66. New York: Holt, Rinehart and Winston. Berlin, Brent, Dennis E. Breedlove, and Peter H. Raven. 1973. General principles of classification and nomenclature in folk biology. American Anthropologist 75:214-242. Berry, J. W., and Pierre R. Dasen, eds. 1974. Culture and cognition. Readings in cross cultural psychology. London: Methuen. Biersack, Aletta. 1982. The logic of misplaced concreteness: Paiela body counting and the nature of the primitive mind. American Anthropologist 84:811-829. Blust, Robert. 1988. Sketches of the morphology and phonology of Bornean languages, 2: Mukah (Melanau). In Papers in Western Austronesian Linguistics, no. 3, ed. Hein Steinhauer, Pacific Linguistics, Series A, No. 78:151-216. Canberra: Australian National University. Bloomfield, Leonard. 1933 Language. London: Allen and Unwin. Bohme, Karin. 1983. Children's understanding and awareness of German possessive pronouns. Doctoral dissertation, Katholieke Universiteit, Nijmegen, The Netherlands.
References
357
Bousfield, John. 1979. The world seen as a colour chart. In Classifications in their social context, eds Roy Ellen and David Reason, 195-220. London: Academic Press. Bright, Jane O., and William Bright. 1969. Semantic structures in Northwestern California and the Sapir-Whorf hypothesis. In Cognitive Anthropology, ed. Stephen Tyler, 66-78. New York: Holt, Rinehart and Winston. Broadfield, A. 1946. The philosophy of classification. London: Graffon and Co. Brown, Roger W., and Eric H. Lenneberg. 1958. Studies in linguistic relativity. In Readings in social psychology, eds. Eleanor Maccoby, Theodore M. Newcomb, and Eugene L. Hartley, 9-18. New York: Holt, Rinehart and Winston. Bruner, Jerome S., Jacqueline J. Goodnow, and George A. Austin. 1962. A study of thinking, with an appendix on language by Roger Brown, New York: Science Editions. Buhler, Karl. 1978 [1934]. Sprachtheorie: Die Darstellungsfunktion der Sprache. Frankfurt am Main, Germany: Ullstein. Burling, Robbins. 1965. How to choose a Burmese numeral classifier. In Context and meaning in cultural anthropology, ed. Melford E. Spiro, 243-264. New York: Free Press. Bussmann, Hadumod. 1983. Lexikon der Sprachwissenschaft. Stuttgart, Germany: Kroner. Capell, Arthur. 1969. A survey of New Guinea languages. Sydney, Australia: Sydney University Press. Capell, Arthur. 1971. The Austronesian languages of Australian New Guinea. In Current Trends in Linguistics, vol. 8: Linguistics in Oceania, ed. Thomas A. Sebeok, 240-340. The Hague: Mouton. Capell, Arthur. 1976. General picture of Austronesian languages, New Guinea area. In New Guinea area languages and language study, vol. 2: Austronesian languages, ed. Stephen A. Wurm, 5-52. Canberra: Australian National University. Carpenter, Kathie. 1986. Productivity and pragmatics of Thai classifiers. In Proceedings of the Twelfth Annual Meeting of the Berkeley Linguistic Society, eds. Vassiliki Nikifaridou, Mary Van Clay, Mary Niepokuj, and Deborah Fedev, 14-25. University of California, Berkeley. Carpenter, Kathie. 1991. Later rather than sooner: Extralinguistic categories in the acquisition of Thai classifiers. Journal of Child Language 18: 93-113. Carpenter, Kathie. 1992. Two dynamic views of classifier systems: Diachronic change and individual development. Cognitive Linguistics 3:129-150. Carroll, John B., and Joseph B. Casagrande. 1958. The function of language classification in behavior. In Readings in social psychology, eds. Eleanor Maccoby, Theodore M. Newcomb, and Eugene L. Hartley, 18-31. New York: Holt, Rinehart and Winston. Chin, Ng Bee. 1989. The acquisition of numeral classifiers in Hokkien, a Southern Min language. Doctoral dissertation, La Trobe University, Bundoora, Victoria, Australia. Cholodovic, A. A. 1954. Ocerk grammatiki korejskogo jazyka. Moscow, Russia, Izdatel'stvo literatury na inostrannych jazykach. Clark, Eve V. 1976. Universal categories: On the semantics of classifiers and children's early word meanings. Linguistic studies offered to Joseph Greenberg: Studia linguistica et philologica, ed. Alphonse Juilland, 449-462. Saratoga, Calif.: Anma Libri. Clark, Herbert, H., Patricia A. Carpenter, and Marcel A. Just. 1973. On the meeting of semantics and perception. In Visual information processing, ed. William G. Chase, 311-381. New York: Academic Press.
358
References
Cole, Michael, John Gay, Joseph A. Glick, and Donald W. Sharp. 1971. The cultural context of learning: An exploration in experimental anthropology. New York, Basic Books. Cole, Michael, and Sylvia Scribner. 1974. Culture and thought: A psychological introduction. New York: John Wiley. Collins, Allan M., and M. R. Quillian. 1969. Retrieval time from semantic memory. Journal of Verbal Learning and Verbal Behavior 8:240-247. Conklin, Harold C. 1962. Lexicographic treatment of folk taxonomies. Problems in lexicography. International Journal of American Linguistics 28(4):119-141. Corum, Claudia, T. Cedric Smith-Stark, and Ann Weiser, eds. 1973. Papers from the Ninth Regional Meeting, Chicago Linguistic Society. Chicago Linguistic Society, University of Chicago, Chicago, I11. Craig, Colette. 1986. Jacaltec noun classifiers: A study in language and culture. In Noun classes and categorization, ed. Colette Craig, 263-293. Amsterdam: John Benjamins. Craig, Colette. 1992. Classifiers in a functional perspective. In Layered structure and reference in a functional perspective, eds. Michael Fortescue, Peter Harder, and Lars Kristoffersen, 277-301. Amsterdam: John Benjamins. Delancey, Scott. 1986. Toward a history of Tai classifier systems. In Noun classes and categorization, ed. Colette Craig, 437-452. Amsterdam: John Benjamins. Demuth, Katherine. 1988. Noun classes and agreement in Sesotho acquisition. In Agreement in natural language: Approaches, theories, descriptions, eds. Michael Barlow and Charles Ferguson, 305-321. Chicago, I11.: University of Chicago Press. Demuth, Katherine. 1992. The acquisition of Sesotho. In The cross-linguistic study of language acquisition, vol. 3, ed. Dan Slobin, 557-638. Hillsdale, N.J.: Lawrence Erlbaum. Demuth, Katherine, Nicolas Faraclas, and Lynell Marchese. 1986. Niger-Congo noun class and agreement systems in language acquisition and historical change. In Noun classes and categorization, ed. Colette Craig, 453-471. Amsterdam: John Benjamins. Denny, J. Peter. 1976. What are noun classifiers good for? In Papers from the Twelfth Regional Meeting, Chicago Linguistic Society, eds. Salikoko S. Mufwene, Carol A. Walker, and Sanford B. Steever, 122-132. Chicago Linguistic Society, University of Chicago, Chicago, I11. Denny, J. Peter. 1979. The "extendedness' variable in classifier semantics: Universal features and cultural variation. In Ethnolinguistics: Boas, Sapir and Whorf revisited, ed. Madeleine Mathiot, 97-119. The Hague: Mouton. Denny, J. Peter. 1986. The semantic role of noun classifiers. In Noun classes and categorization, ed. Colette Craig, 297-308. Amsterdam: John Benjamins. Diller, A. V. N. 1985. High and low Thai: Views from within. In Papers in South-East Asian Linguistics, ed. David Bradley, 51-76. Canberra: Australian National University. Dixon, Robert M. W. 1968. Noun classes. Lingua 21:104-125. Dixon, Robert M. W. 1972. The Dyirbal language of North Queensland. Cambridge: Cambridge University Press. Dixon, Robert M. W. 1982. Where have all the adjectives gone and other essays in semantics and syntax. Berlin: Mouton. Dixon, Robert M. W. 1986. Noun classes and noun classification in typological perspective. In Noun classes and categorization, ed. Colette Craig, 105-112. Amsterdam: John Benjamins. Dolby, R. G. A. 1979. Classification of the sciences. The nineteenth century tradition.
References
359
In Classifications in their social context, eds. Roy F. Ellen and David Reason, 167-193. London: Academic Press. Downing, Pamela. 1986. The anaphoric use of classifiers in Japanese. Noun classes and categorization, ed. Colette Craig, 345-375. Amsterdam: John Benjamins. Durkheim, Emile, and Marcel Mauss. 1970 [1903]. Primitive classification. Translated by Rodney Needham. London: Routledge and Kegan Paul. (De quelques formes primitives de classification. Annee Sociologique 1901-02.) Eibl-Eibesfeldt, Irenaus. 1984. Die Biologie des menschlichen Sozialverhaltens. GrundriB der Humanethologie. Munchen, Germany: Piper. Eibl-Eibesfeldt, Irenaus, Barbara Senft, and Gunter Senft. 1987. Trobriander (Ost-Neuguinea, Trobriand Inseln, Kaile'una) Fadenspiele "ninikula". In Publikationen zu Wissenschaftlichen Filmen, Sektion Ethnologie, Serie 15, Nummer 25. Gottingen, Germany: Institut fur den Wissenschaftlichen Film. Eibl-Eibesfeldt, Irenaus, and Gunter Senft. 1986. Trobriander (Ost-Neuguinea, Trobriand Inseln, Kaile'una) Fadenspiele "ninikula". Film Nr. E 2958. Gottingen, Germany: Institut fur den Wissenschaftlichen Film. Ellen, Roy F. 1979. Introductory Essay. In Classifications in their social context, eds. Roy F. Ellen and David Reason, 1-32. London: Academic Press. Ellen, Roy F., and David Reason. 1979. Preface. In Classifications in Their Social Context, eds. Roy F. Ellen and David Reason, vii-viii. London: Academic Press. Erbaugh, Mary S. 1986. Taking stock: The development of Chinese noun classifiers historically and in young Children. In Noun classes and categorization, ed. Colette Craig, 399—436. Amsterdam: John Benjamins. Ezard, Bryan. 1978. Classificatory prefixes of the Massim cluster. In Second International Conference on Austronesian Linguistics: Proceedings, Fascicle 2, eds. Stephen A. Wurm and Lois Carrington, 1159-1180. Pacific Linguistics, Series C, No. 61. Canberra: Australian National University. Fellows, S. B. 1901. Grammar of the Kiriwina dialect (together with a vocabulary). In Annual Report on British New Guinea 1900-1901, appendix N2, 171-177; appendix O2, 177-196. Brisbane, Australia: G. A. Vaughan. Fischer, Gero. 1972. Die syntaktischen und semantischen Eigenschaften der Numeralklassifikatoren im Thai. Archiv Orientalni 40:65-78. Foley, William A. 1968. The Papuan languages of New Guinea. Cambridge: Cambridge University Press. Foucault, Michel. 1980 [1966]. Die Ordnung der Dinge. Eine Archaologie der Humanwissenschaften. Frankfurt am Main, Germany: Suhrkamp. (Les Mots et les choses. Paris: Gallimard.) Fox, James J. 1975. On binary categories and primary symbols. Some Rotinese perspectives. In The interpretation of symbolism, ed. Roy Willis, 99-132. London: Malaby Press. Fox, James J. 1989. Category and complement: Binary ideologies and the organization of dualism in Eastern Indonesia. In The attraction of opposites: Thought and society in the dualistic mode, eds. David Maybury-Lewis and Uri Almagar, 33-56. Ann Arbor: University of Michigan Press. Frake, Charles O. 1969. The ethnographic study of cognitive systems. In Cognitive anthropology, ed. Stephen A. Tyler, 28-41. New York: Holt, Rinehart and Winston. Freilich, Merris, ed. 1970. Marginal natives: Anthropologists at work. New York: Harper and Row. Friedrich, Paul. 1970. Shape in grammar. Language 46:379-407. Gandour, Jack, Saranee H. Petty, Rochama Dardarananda, Suralee Dechongkit, and
360
References
Sunee Mukngoen. 1984. The acquisition of numeral classifiers in Thai. Linguistics 22:455-479. Gipper, Helmut. 1972. Gibt es ein sprachliches Relativitatsprinzip? Untersuchungen zur Sapir-Whorf-Hypothese. Frankfurt am Main, Germany: S. Fischer. Givon, Talmy. 1986. Prototypes: Between Plato and Wittgenstein. In Noun classes and categorization, ed. Colette Craig, 77-102. Amsterdam: John Benjamins. Goffman, Erving. 1974. Frame analysis. An essay on the organization of experience. New York: Harper and Row. Gohlke, Paul, ed. 1972. Aristoteles—Kategorien und Hermeneutik. Paderborn, Germany: Schoningh. Gomez-Imbert, Elsa. 1982. De la forme et du sens dans la classification nominale in Tatuyo (langue Tukano Orientale d'Amazonie Colombienne). Doctoral dissertation, Universite Sorbonne, Paris IV, Ecole Pratique des Hautes Etudes, IVe section. Goodenough, Ward H. 1969. Rethinking "status" and "role": Toward a general model of the cultural organization of social relationships. In Cognitive Anthropology, ed. Stephen Tyler, 311-330. New York: Holt, Rinehart and Winston. Goral, Donald A. 1978. Numeral classifier systems: A South-East Asian cross-linguistic analysis. Linguistics in the Tibeto-Burman Area 4:1-72. Greenberg, Joseph H. 1972. Numerical classifiers and substantival number: Problems in the genesis of a linguistic type. Working Papers in Language Universals 9:1-39. Greenberg, Joseph H. 1975. Numeral classifiers and substantival number: Problems in the genesis of a linguistic type. In Proceedings of the l1th International Congress of Linguistics, ed. Luigi Heilmann, 17-37. Bologna, Italy: Mulino. Greenberg, Joseph H. 1978. How does a language acquire gender-markers? In Universals of human language, vol. 3, Word structure, ed. Joseph H. Greenberg, 47-82. Stanford, Calif.: Stanford University Press. Haas, Mary R. 1942. The use of numeral classifiers in Thai. Language 18:201-205. Hallpike, Christopher R. 1979. The foundations of primitive thought. Oxford: Clarendon Press. Harweg, Roland 1987a. Stoffhamen und Gattungsnamen. Zeitschrift fur Phonetik, Sprachwissenschaft und Kommunikationsforschung 40:792-804. Harweg, Roland. 1987b. Zahleinheitsworter und Weltbild. Archiv Orientalni 55:253-270. Haugen, Einar. 1977. Linguistic relativity: Myths and methods. In Language and thought: Anthropological issues, ed. William McCormack and Stephen A. Wurm, 11-28. The Hague: Mouton. Heeschen, Volker. 1972. Die Sprachphilosophie Wilhelm von Humboldts. Doctoral dissertation, Ruhr-Universitat, Bochum, Germany. Heine, Bernd. 1982. African noun-class systems. In Apprehension: das sprachliche Erfassen von Gegenstanden, Teil 1: Bereich und Ordnung der Phdnomene, ed. Hansjakob Seiler and Christian Lehmann, 189-216. Tubingen, Germany: Narr. Herder, Johan Gottfried 1978 [1770]. liber den Ursprung der Sprache. In Herders Werke in funf Banden. Band 2, 91-200. Berlin: Aufbau Verlag. Hiranburana, SamAng. 1979. A classification of Thai classifiers. South-East Asian Linguistic Studies, vol. 4, ed. Dang Liem Nguyen, 39-54. (Pacific Linguistics, Series C, No. 49.) Canberra: Australian National University. Hla Pe. 1965. A Re-examination of Burmese "classifiers". Lingua 15:163-185. Hoa, Ngyen Dinh. 1957. Classifiers in Vietnamese. Word 13:124-152.
References
361
Hopper, Paul J. 1986. Some discourse functions of classifiers in Malay. In Noun classes and categorization, ed. Colette Craig, 309-325. Amsterdam: John Benjamins. Humboldt, Wilhelm von. 1968 [1836]. Uber die Verschiedenheit des menschlichen Sprachbaues und ihren EinfluB auf die geistige Entwicklung des Menschengeschlechts. Bonn, Germany: Dummler. Hundius, Harald, and Ulrike Kolver. 1983. Syntax and semantics of numeral classifiers in Thai. Studies in Language 7:165-214. Hutchins, Edwin. 1980. Culture and inference. A Trobriand case study. Cambridge, Mass.: Harvard University Press. Jackendoff, Ray S. 1968. Quantifiers in English. Foundations of Language 4:422-442. Jacob, Judith M. 1965. Notes on numerals and numeral co-efficients in Old, Middle and Modern Khmer. Lingua 15:143-162. Jensen, Hans. 1952. Die sprachliche Kategorie des Numerus. Wissenschaftliche Zeitschrift der Universitat Rostock 1(H3):1-21. Jones, Robert B. 1970. Classifier constructions in Southeast Asia. Journal of the American Oriental Society 90:1-12. Juilland Alphonse, and Hans Heinrich Lieb. 1968. Klasse und Klassifikation in der Sprachwissenschaft. The Hague: Mouton. Kaden, Klaus. 1964. Der Ausdruck der Mehrzahlverhaltnisse in der modernen chinesischen Sprache. Berlin: Akademie Verlag. Kantor, Rebecca. 1980. The acquisition of classifiers in American Sign Language. Sign Language Studies 28:193-208. Kasaipwalova, John. 1980. Sail the midnight sun. Port Moresby, Papua New Guinea: National Arts School. Katz, Elisabeth. 1982. Zur Distribution von Kompositum und Nominalgruppe im Deutschen. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 1: Bereich und Ordnung der Phanomene, ed. Hansjakob Seiler and Christian Lehmann, 112-129. Tubingen, Germany: Narr. Kegl, Judy, and Sara Schley. 1986. When is a classifier no longer a classifier? In Proceedings of the 12th Annual Meeting of the Berkeley Linguistic Society, 425-441. Berkeley, Calif.: Berkeley Linguistic Society. Kendon, Adam. 1988. Sign languages of Aboriginal Australia: Cultural, semiotic and communicative perspectives. Cambridge: Cambridge University Press. Kessel, Frank S., ed. 1988. The development of language and language researchers: Essays in honor of Roger Brown. Hillsdale, N.J.: Lawrence Erlbaum. Klaus, Georg, ed. 1968. Worterbuch der Kybernetik. Berlin: Dietz. Klein, Harriet E. Manelis. 1979. Noun classifiers in Toba. In Ethnolinguistics: Boas, Sapir and Whorf revisited, ed. Madeleine Mathiot, 85-95. The Hague: Mouton. Klein, Wolfgang. 1974. Variation in der Sprache. Kronberg, Germany: Scriptor. Klein, Wolfgang. 1984. Zweitspracherwerb. Eine Einfuhrung. Konigstein, Germany: Athenaum. Klein, Wolfgang. 1986. Second language acquisition. Cambridge: Cambridge University Press. Klein, Wolfgang, and Norbert Dittmar. 1979. Developing grammars. Berlin: Springer. Klima, Edward S., and Ursula Bellugi. 1979. The signs of language. Cambridge, Mass.: Harvard University Press. Koch, Gerd. 1984. Malingdam. Mensch, Kultur und Umwelt im zentralen Bergland von West-Neuguinea, vol. 15. Berlin: Reimer. Koch, Walter A. 1986. Genes vs. memes: Modes of integration for natural and cultural evolution in a holistic model ("ELPIS"). Bochum, Germany: Brockmeyer.
362
References
Koestler, Arthur. 1983. Janus: A summing up. London: Pan Picador. Kolver, Ulrike. 1978. On Newari noun phrases. In Language universals, ed. Hansjakob Seiler, 277-300. Tubingen, Germany: Narr. Kolver, Ulrike. 1979. Syntaktische Untersuchung von Numeralklassifikatoren im Zentralthai. Akup, 34, ed. Hansjakob Seiler. Cologne, Germany: Universitat zu Koln. Kolver, Ulrike. 1982a. Zu den Techniken Numerus, Kollektion und Numeralklassifikation in der Bahasa Indonesia. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 2: Die Techniken und ihr Zusammenhang in Einzelsprachen, ed. Hansjakob Seiler and Franz-Josef Stachowiak, 107-122. Tubingen, Germany: Narr. Kolver, Ulrike. 1982b. Interaktion von nominalen Kategorien am Beispiel der Entwicklung des modernen Bengali. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 1: Bereich und Ordnung der Phanomene, ed. Hansjakob Seiler and Christian Lehmann, 244-251. Tubingen, Germany: Narr. Kolver, Ulrike. 1982c. Klassifikatorkonstruktion in Thai, Vietnamesisch und Chinesisch—Ein Beitrag zur Dimension der Apprehension. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 1: Bereich und Ordnung der Phanomene, ed. Hansjakob Seiler and Christian Lehmann, 160-185. Tubingen, Germany: Narr. Kolver, Ulrike. 1983. The notion of scalarity in language comparison. In Language invariants and mental operations, ed. Hansjakob Seiler and Gunter Brettschneider, 53-57. Tubingen, Germany: Narr. Koschmieder, Erwin. 1945. Zur Bestimmung der Funktion grammatischer Kategorien. In Abhandlungen der Bayerischen Akademie der Wissenschaften: Philosophischhistorische Abteilung, NF25:l-64. Labov, William. 1972a. Language in the inner city: Studies in the black English vernacular. Philadelphia: University of Pennsylvania Press. Labov, William. 1972b. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. Lakoff, George. 1986. Classifiers as a reflection of mind. Noun classes and categorization, ed. Colette Craig, 13-51. Amsterdam: John Benjamins. Lakoff, George. 1987. Women, fire, and dangerous things. What categories reveal about the mind. Chicago: University of Chicago Press. Lawton, Ralph. 1980. The Kiriwinan classifiers. Masters thesis. Canberra: Australian National University. Leach, Edmund. 1972 (1964). Anthropologische Aspekte der Sprache: Tierkategorie und Schimpfworter. In Biologische Grundlagen der Sprache, ed. Eric Lenneberg, 32-73. Frankfurt am Main, Germany: Suhrkamp (Biological Foundations of Language. New York: John Wiley). Leach, Jerry W. 1949/50. Primitive calendars. Oceania 20:254-262. Leach, Jerry W. 1981. A Kula folktale from Kiriwina. Bikmaus 2(l):50-92. Lean, G. A. 1986. Counting systems of Papua New Guinea—Research bibliography. Lae, Papua New Guinea: Mathematics Education Centre, PNG University of Technology. Lee, Michael. 1987. The cognitive basis of classifier systems. Berkeley Linguistic Society, Proceedings of the 13th Annual Meeting, 395-407. Berkeley, Calif.: Berkeley Linguistic Society. Lee, Michael. 1988. Language, perception and the world. In Explaining language universals, ed. John Hawkins, 211-246. Oxford: Basil Blackwell.
References
363
Lehmann, F. K. 1979. Aspects of a formal theory of classifiers. Studies in Language 3:153-180. Lenneberg, Eric H. 1953. Cognition in ethnolinguistics. Language 29:463-471. Lenneberg, Eric H., ed. 1972 Neue Perspektiven in der Erforschung der Sprache. Frankfurt am Main, Germany: Suhrkamp (New Directions in the Study of Language. Cambridge, Mass.: MIT Press). Lenneberg, Eric H. 1977 (1967). Biologische Grundlagen der Sprache. Frankfurt am Main: Suhrkamp (Biological Foundations of Language. New York: John Wiley), de Leon, Lourdes. 1988. Noun and numeral classifiers in Mixtec and Tzotzil: A referential view. Doctoral dissertation, University of Sussex, Linguistics Division of the Graduate School, Sussex, England. Lenz, Jakob Michael Reinhold. 1909 [1774]. Der neue Menoza oder Geschichte des kumbanischen Prinzen Tandi. Eine Komodie. In Jakob Michael Reinhold Lenz, Gesammelte Schriften. Herausgegeben von Franz Blei, vol. 2. Munich, Germany: Georg Muller. Levelt, Willem J. M. 1980. On-line processing constraints on the properties of signed and spoken language. In Signed and spoken language: Biological constraints on linguistic forms, ed. Ursula Bellugi and Michael Studdert-Kennedy, 141-160. Weinheim, Germany: Verlag Chemie. Levinson, Stephen C. 1983. Pragmatics. Cambridge: Cambridge University Press. Lithgow, David. 1976. Austronesian languages: Milne Bay Province and adjacent islands (Milne Bay Province). In New Guinea area languages and language study, vol 2: Austronesian Languages, ed. Stephen A. Wurm, 441-523. Pacific Linguistics, Series C, No. 39. Canberra: Australian National University. Lobel, Elisabeth. 1986. Apposition in der Quantifizierung. In Pragmantax, eds. Armin Burkhardt and Karl-Hermann Korner, 47-59. Tubingen, Germany: Niemeyer. Lonning, Jam Tare. 1987. Mass terms and quantification. Linguistics and Philosophy 10:1-52. Lorenz, Konrad. 1973. Die Ruckseite des Spiegels: Versuch einer Naturgeschichte menschlichen Erkennens. Munich, Germany: Piper. Lorenz, Konrad. 1977: Behind the mirror: A search for a natural history of human knowledge. Translated by Ronald Taylor. New York: Harcourt Brace Jovanovich. Luke, K. K., and Godfrey Harrison, 1986. Young children's use of Chinese (Cantonese and Mandarin) sortal classifiers. In Linguistics, psychology and the Chinese language, eds. S. R. Henry Kao and Rumjahn Hoosain, 125-147. Centre of Asian Studies. Hong Kong: University of Hong Kong. Lyons, John. 1977a. Semantics I. Cambridge: Cambridge University Press. Lyons, John. 1977b. Semantics II. Cambridge: Cambridge University Press. MacLaury, Robert E. 1991. Prototypes revisited. Annual Review of Anthropology 20:55-74. Malinowski, Bronislaw. 1920. Classificatory particles in the language of Kiriwina. Bulletin of the School of Oriental Studies I(IV):33-78. Malinowski, Bronislaw. 1978 [1922]. Argonauts of the Western Pacific. London: Routledge and Kegan Paul. Malinowski, Bronislaw. 1982 [1929]. The sexual life of savages in North Western Melanesia. London: Routledge and Kegan Paul. Malinowski, Bronislaw. 1966 [1935]. Coral gardens and their magic, 2 vols. London: Allen and Unwin. Malinowski, Bronislaw. 1974. Magic, science and religion and other essays. London: Souvenir Press.
364
References
Malmberg, Bertil. 1987. Wilhelm von Humboldt und die moderne Sprachwissenschaft. In Preprints of the Plenary Session Papers of the XlVth International Congress of Linguists organized under the Auspices of CIPL, Berlin, 10-15. August 1987, 3-25. Berlin: Editorial Committee. Martens, Michael. 1988. How big is your rice? Units of measurement among the Uma people. Papers in Western Austronesian Linguistics, No. 4, 257-261. Pacific Linguistics, Series A. No. 79. Canberra: Australian National University. Matsumoto, Yo. 1985. Acquisition of some Japanese numeral classifiers: The search for convention. Papers and Reports in Child Language Development, Stanford University, 24:79-86. Miller, George A. 1956. The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review 63(2):81-96. Miller, George A., and Philip N. Johnson-Laird. 1976. Language and perception. Cambridge, Mass.: Harvard University Press. Milner, George B. 1963. Oceanic linguistics. In Trends in modern linguistics, eds. Christine Mohrmann, F. Norman, and Alf Sommerfelt, 62-94. Utrecht, The Netherlands: Spectrum Publishers. Miram, Helga-Maria. 1983. Numeral classifiers im yukatekischen Maya. Hannover, Germany: Verlag fur Ethnologie. Mithun, Marianne. 1986. The convergence of noun classification systems. In Noun classes and categorization, ed. Colette Craig, 379-397. Amsterdam: John Benjamins. Moerman, Michael. 1969. A little knowledge. In Cognitive Anthropology, ed. Stephen A. Tyler, 449-469. New York: Holt, Rinehart and Winston. Mufwene, Salikoko S. 1980a. Number, countability and markedness in Lingala LI-/MA-noun class. Linguistics 18:1019-1052. Mufwene, Salikoko S. 1980b. "Prototype" and "Kin-Class". Anthropological Linguistics 22:29-41. Mufwene, Salikoko S. 1980c. Bantu class prefixes: Inflectional or derivational? In Papers from the Sixteenth Regional Meeting, Chicago Linguistic Society, 246-258. Chicago Linguistic Society, University of Chicago, Chicago, I11. Mufwene, Salikoko S. 1981. Noun individuation and the count/mass distinction. In Papers from the Seventeenth Regional Meeting, Chicago Linguistic Society, 221-238. Chicago Linguistic Society, University of Chicago, Chicago, I11. Mufwene, Salikoko S. 1983. Investigating what the words Father and Mother mean. Language and Communication 3:245-269. Mufwene, Salikoko S. 1984. The count/mass distinction and the English lexicon. In Papers from the parasession on lexical semantics, Chicago Linguistic Society, 200-221. Chicago Linguistic Society, University of Chicago, Chicago, I11. Newport, Elissa, and Ted Supalla. 1980. Clues from the acquisition of signed and spoken language. In Signed and spoken language: Biological constraints on linguistic form, eds Ursula Bellugi and Michael Studdert-Kennedy, 187-212. Weinheim, Germany: Verlag Chemie. Ohta, Itaru. 1987. Livestock individual identification among the Turkana: The animal classification and naming in the pastoral livestock management. In African Study Monographs, vol. 8, no. 1: 1-69. Kyoto, Japan: Kyoto University. Parsons, Terence. 1970. An analysis of mass terms and amount terms. Foundations of Language 6: 362-388. Plank, Frans. 1984. Verbs and objects in semantic agreement: Minor differences between English and German that might suggest a major one. Journal of Semantics 3:305-360.
References
365
Posner, Michael. 1986. Empirical studies on prototypes. In Noun classes and categorization, ed. Colette Craig, 53-61. Amsterdam: John Benjamins. Powell, Harry A. 1957. An analysis of present day social structure in the Trobriand Islands. Doctoral dissertation, University of London, London, England. Powell, Harry A. 1960. Competitive leadership in Trobriand political organization. Journal of the Royal Anthropological Institute 90:118-148. Pribram, Karl H. 1987. Brain and meaning. Mimeo, Stanford University, Stanford, Calif. Pullum, Geoffrey K. 1991. The great Eskimo vocabulary hoax and other irrelevant essays on the study of language. Chicago: University of Chicago Press. Raible, Wolfgang. 1981. Von der Allgegenwart des Gegensinns (und einiger anderer Relationen): Strategien zur Einordnung semantischer Information. Zeitschrift fur romanische Philologie 97:1-40. Rausch, P. J. 1912. Die Sprache von Sudost-Bougainville, Deutsche Salomonsinseln. Anthropos 7:105-134, 585-616, 964-994. Ray, Sidney H. 1907. Reports of the Cambridge Anthropological Expedition to Torres Straits, vol. 3: Linguistics. Cambridge: Cambridge University Press. Ray, Sidney H. 1938. The languages of the Eastern Lousiade Archipelago. Bulletin of the School of Oriental Studies 9:363-384. Romaine, Suzanne. 1985. Variable rules, O.K.? Or can there be sociolinguistic grammars? Language and Communication 5:53-67. Rosch, Eleanor. 1977. Human categorization. Studies in Cross-Cultural Psychology, vol. 1, ed. Neil Warren, 1-49. London: Academic Press. Rosch, Eleanor. 1978. Principles of categorization. In Cognition and Categorization, eds. Eleanor Rosch and Barbara B. Lloyd, 27-48. Hillsdale, N.J.: Lawrence Erlbaum. Rosch, Eleanor. 1988. Coherence and categorization: A historical view. In The development of language and language researchers: Essays in honor of Roger Brown, ed. Frank S. Kessel, 373-392. Hillsdale, N.J.: Lawrence Erlbaum. Ross, Malcolm D. 1988. Proto Oceanic and the Austronesian Languages of Western Melanesia. Pacific Linguistics, Series C, No. 98. Canberra: Australian National University. Royen, Gerlach. 1929. Die nominalen Klassifikations-Systeme in den Sprachen der Erde: Historisch-kritische Studie, mit besonderer Berucksichtigung des Indogermanischen. Anthropos Linguistische Bibliothek, vol. IV. Vienna: Anthropos. Rude, Noel. 1986. Graphemic classifiers in Egyptian hieroglyphics and Mesopotamian cuneiform. In Noun classes and categorization, ed. Colette Craig, 133-138. Amsterdam: John Benjamins. Sanches, Mary 1977. Language acquisition and language change: Japanese numeral classifiers. Sociocultural dimensions of language change, eds. Ben G. Blount and Mary Sanches, 51-62. New York: Academic Press. Saul, Janice E. 1965. Classifiers in Nung. Lingua 13:278-290. Schafer, Edward H., Jr. 1948. Noun classifiers in classical Chinese. Language 24:408-413. Schiefenhovel, Wulf, and Ingrid Bell-Krannhals. 1986. Wer teilt hat teil an der Macht: Systeme der Yams-Vergabe auf den Trobriand Inseln, Papua-Neuguinea. Mitteilungen der Anthropologischen Gesellschaft in Wien 116:19-39. Schieffelin, Bambi B., and Elinor Ochs. 1986. Language socialization. Annual Review of Anthropology 15:163-191. Schleiermacher, Friedrich D. E. 1977 [1838]. Hermeneutik und Kritik mit besonderer Beziehung auf das neue Testament. Aus Schleirmachers handschriftlichem Nachlasse
366
References
und nachgeschriebenen Vorlesungen herausgegeben und eingeleitet von Manfred Frank. Frankfurt am Main, Germany: Suhrkamp. Schmidt, Arno. 1975. Ansichten aus dem Leben eines Lords. Frankfurt am Main, Germany: Fischer. Schone, Hermann. 1983. Orientierung im Raum: Formen und Mechanismen der Lenkung des Verhaltens im Raum bei Tier und Mensch. Stuttgart, Germany: Wissenschaftliche Verlagsanstalt. Scoditti, Giancarlo M. G. 1985. Kitawa: Iconogrqfia e semantica in una societa melanesiana. Milan, Italy: Franco Angeli Libri. Seiler, Hansjakob. 1973. Das Universalienkonzept. In Linguistic Workshop I: Vorarbeiten zu einem Universalienprojekt, ed. Hansjakob Seiler. Mimeo, Munich, Germany. Seiler, Hansjakob. 1978. Language Universals. Tubingen, Germany: Narr. Seiler, Hansjakob. 1982. Das sprachliche Erfassen von Gegenstanden Apprehension. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 1: Bereich und Ordnung der Phdnomene, eds. Hansjakob Seiler and Christian Lehmann, 3-11. Tubingen, Germany: Narr. Seiler, Hansjakob. 1986. Apprehension. Language, object, and order, part III: The universal dimension of apprehension. Tubingen, Germany: Narr. Seiler, Walter. 1983. The acquisition of a noun classification system by a language. Language and Linguistics in Melanesia 14:76-97. Seiler, Walter. 1986. From verb serialisation to noun classification. In Papers in New Guinea Linguistics, No. 24, 11-19. Pacific Linguistics, Series A, No. 70. Canberra: Australian National University. Senft, Barbara. 1985. Kindheit in Tauwema. Die ersten 7 Jahre im Leben der Kinder auf den Trobriand-Inseln, Papua-Neuguinea. Mimeo, Seewiesen, Germany: Forschungsstelle fur Humanethologie. Senft, Barbara, and Gunter Senft. 1986. Ninikula—Fadenspiele auf den Trobriand Inseln, Papua Neuguinea. Untersuchungen zum Spiele-Repertoire unter besonderer Berucksichtigung der Spiel-begleitenden Texte. Baessler-Archiv, Beitrage zur Volkerkunde NF 34(l):93-235. Senft, Gunter. 1982. Sprachliche Varietat und Variation im Sprachverhalten Kaiserslauterer Metallarbeiter: Untersuchungen zu ihrer Begrenzung, Beschreibung, und Bewertung. Bern, Switzerland: Peter Lang. Senft, Gunter. 1983. The system of classificatory particles in Kilivila reconsidered—First results on its inventory, its acquisition, and its usage. Paper presented at the Annual Meeting of the Linguistic Society of Papua New Guinea, Port Moresby, University of PNG, July 1983. Senft, Gunter. 1985a. Klassifikationspartikel im Kilivila—Glossen zu ihrer morphologischen Rolle, ihrem Inventar und ihrer Funktion in Satz und Diskurs. Linguistische Berichte 99:373-393. Senft, Gunter. 1985b. Kilivila—die Sprache der Trobriander. Studium Linguistik 17/18:127-138. Senft, Gunter. 1985c. Weyeis Wettermagie—eine ethnolinguistische Untersuchung von funf wettermagischen Formeln eines Wettermagiers auf den Trobriand Inseln. Zeitschrift fur Ethnologie 10(1):67-90; (H2):erratum. Senft, Gunter. 1985d. Trauer auf Trobriand—Eine ethnologisch/linguistische Fallstudie. Anthropos 80:471-492. Senft, Gunter. 1985e. How to tell—and understand—a "dirty" joke in Kilivila. Journal of Pragmatics 9:815-834.
References
367
Senft, Gunter. 1985f. Emic or etic or just another catch 22? A repartee to Hartmut Haberland. Journal of Pragmatics 9:845. Senft, Gunter. 1986. Kilivila. The language of the Trobriand Islanders. Mouton Grammar Library, 3. Berlin: Mouton de Gruyter. Senft, Gunter. 1987a. The system of classiflcatory particles in Kilivila reconsidered— First results on its inventory, its acquisition, and its usage. Language and Linguistics in Melanesia 16:100-125. Senft, Gunter. 1987b. Kilivila color terms. Studies in Language 11:315-346. Senft, Gunter. 1987c. Rituelle Kommunikation auf den Trobriand Inseln. Zeitschrift fur Literaturwissenschaft und Linguistik 65:105-130. Senft, Gunter. 1987d. Nanam'sa Bwena—Gutes Denken: Eine ethnolinguistische Fallstudie uber eine Dorfversammlung auf den Trobriand Inseln. Zeitschrift fur Ethnologie 112:181-122. Senft, Gunter. 1989. Apropos the whole and its parts: Classificatory particles in Kilivila language. In Das Ganze und seine Teile—The Whole and its Parts, ed. Walter A. Koch, 142-176. Bochum, Germany: Brockmeyer. Senft, Gunter. 199la. Bakavilisi Biga—We can "turn" the language—Or: What happens to English words in Kilivila language? In Proceedings of the XIVth International Congress of Linguists (held in Berlin/GDR, 1987). Organized under the Auspices of CIPL, eds Werner Bahner, Joachim Schildt; and Dieter Viehweger, vol. II, 1743-1746. Berlin: Akademie Verlag. Senft, Gunter. 1991b. Prolegomena to the pragmatics of "situational-intentional" varieties in Kilivila language. In Levels of linguistic adaptation: Selected papers from the 1987 International Pragmatics Conference, vol. II, ed. Jef Verschueren, 235-248. Amsterdam: John Benjamins. Senft, Gunter. 1991c. Network models to describe the Kilivila classifier system. Oceanic Linguistics 30:131-155. Senft, Gunter. 1991d. Mahnreden auf den Trobriand Inseln—Eine Fallstudie. Verbale Interaktion—Studien zur Empirie und Methodologie der Pragmatik, ed. Dieter Flader, 27-49. Stuttgart, Germany: J. B. Metzler. Senft, Gunter. 1992a. Bakavilisi Biga—Or: What happens to English words in the Kilivila language? Language and Linguistics in Melanesia 23:13-49. Senft, Gunter. 1992b. What happened to "the fearless tailor" in Kilivila: A European fairy tale from the South Seas. Anthropos 87:407-421. Senft, Gunter. 1992c. Body and mind in the Trobriand Islands. Paper presented at the 91st Annual Meeting of the American Anthropological Association, San Francisco, December 2-6, 1992. Mimeo, Nijmegen, Max-Planck-Institute, The Netherlands. Senft, Gunter. 1993. A grammaticalization hypothesis on the origin of Kilivila classificatory particles. Sprachtypologie und Universalienforschung 46:100-112. Senft, Gunter. In press. Zeichenkonzeptionen in Ozeanien. Semiotics: A handbook on the sign-theoretic foundation of nature and culture, eds. Roland Posner, Klaus Robering, and Thomas A. Sebeok. Berlin: de Gruyter. Serzisko, Fritz. 1980. Sprachen mit Zahlklassifikatoren Analyse und Vergleich. In Akup Nr. 37, ed. Hansjakob Seiler. Cologne, Germany: Universitat zu Koln. Serzisko, Fritz. 1982a. Temporare Klassifikation. Ihre Variationsbreite in Sprachen mit Zahlklassifikatoren. In Apprehension: das sprachliche Erfassen von Gegenstanden, part 1: Bereich und Ordnung der Phanomene, eds. Hansjakob Seiler and Christian Lehmann, 147-159. Tubingen, Germany: Narr. Serzisko, Fritz. 1982b. Gender, noun class, and numeral classification: a scale of classificatory techniques. In Issues in the theory of universal grammar, eds. Rene Dirven and Gunter Radden, 95-123. Tubingen, Germany: Narr.
368
References
Silverstein, Michael. 1986. Classifiers, verb classifiers, and verbal categories. In Proceedings of the 12th Annual Meeting of the Berkeley Linguistic Society, 497-514. Berkeley, Calif.: Berkeley Linguistic Society. Sinclar, Anne, Robert J. Jarvella, and Willem J. M. Levelt. 1978. The child's conception of language. Heidelberg, Germany: Springer. Slobin, Dan I. 1991. Learning to think for speaking: Native language, cognition, and rhetorical style. Pragmatics 1:7-25. Smith, Kenneth D. 1979. Sedang Grammar. Pacific Linguistics, Series B, No. 50. Canberra: Australian National University. Spitulnik, Debra. 1988. Levels of semantic structuring in Bantu noun classification. In Current approaches to African linguistics, vol. 5, eds. Paul Newman and Robert D. Botne, 207-220. Dordrecht, The Netherlands: Foris. Stern, Clara, and William Stern. 1981 [1928]. Die Kindersprache. Darmstadt, Germany: Wissenschaftliche Buchgesellschaft. Stolz, Thomas 1991. Die aztekischen Numeralklassifikatoren. Mimeo, Ruhr-University Bochum, Germany. Strawson, P. F. 1950. On referring. Mind 59:320-344. Supalla, Ted. 1986. The classifier system in American Sign Language: Numeral classifiers in Austroasiatic. In noun classes and categorization, ed. Colette Craig, 181-214. Amsterdam: John Benjamins. Tversky, Barbara. 1986. Components and categorization. Noun classes and categorization, ed. Colette Craig, 63-75. Amsterdam: John Benjamins. Tyler, Stephen A. 1969. Introduction. Cognitive anthropology, ed. Stephen A. Tyler, 1-23. New York: Holt, Rinehart and Winston. Unterbeck, Barbara. 1990a. Thesen zur Dissertation: Kollektion, Numeralklassifikation und Transnumerus— Uberlegungen zur Apprehension im Koreanischen und zu einer typologischen Charakteristik von Substantiven. Zentralinstitut fur Sprachwissenschaft, Akademie der Wissenschaften der DDR, Berlin. Unterbeck, Barbara. 1990b. Kollektion, Numeralklassifikation und Transnumerus— Uberlegungen zur Apprehension im Koreanischen und zu einer typologischen Charakteristik von Substantiven. Doctoral dissertation. Zentralinstitut fur Sprachwissenschaft, Akademie der Wissenschaften der DDR. Vachek, Joseph. 1976. Linguistic typology and linguistic characterology. In Linguistic studies offered to Joseph Greenberg: Studia Linguistica et Philologica, ed. Alphonse Juilland, 223-226. Saratoga, Calif.: Anma Libri. Vater, Heinz. 1991. Einfuhrung in die Raum-Linguistik, Klage 24. Cologne, Germany: Gabel Verlag. Verhaar, John. 1986. Information structure and noun phrases. Sophia Linguistica: Working Papers in Linguistics 20/21:1-6. Verhaar, John. 1987. Book review of Morton Benson et al. (1986). The BBI combinatory dictionary of English: A guide to word combination. Studies in Language 11:503-504. Verhaar, John. 1988. Phrase syntax in contemporary Indonesian: Noun phrases. In Towards a description of contemporary Indonesian: Preliminary studies, part III, ed. Bambang Kaswanti Purwo, 1—45. Jakarta: Nusa. Vollmer, Gerhard. 1988a. Was konnen wir wissen? vol. 1: Die Natur der Erkenntnis. Stuttgart, Germany: Hirzel. Vollmer, Gerhard. 1988b. Was konnen wir wissen? vol. 2: Die Erkenntnis der Natur. Stuttgart, Germany: Hirzel. Vollmer, Gerhard. 1989. Auf der Suche nach der Ordnung—Naturphilosophische Uberlegungen zur Entwicklung von Weltbildern. In Das Ganze und seine Teile—The Whole and its Parts, ed. Walter A. Koch, 194-221. Bochum, Germany: Brockmeyer.
References
369
Wallace, Anthony F. C. 1962. Culture and cognition. Science 135:351-357. Wallace, Ron. 1989. Cognitive mapping and the origin of language and mind. Current Anthropology 30:518-526. Wassmann, Jurg, and Pierre R. Dasen. 1994. Yupno Number System and Counting. Journal of Cross-Cultural Psychology 25(2):78-94. Watson-Gegeo, Karen Ann, and David W. Gegeo. 1986. The social world of Kwara'ae children: Acquisition of language and values. In Children's world and children's language, eds. Jenny Cook-Gumperz, William Corsaro, and Jurgen Streeck, 109-127. Berlin: Mouton de Gruyter. Weiner, Annette B. 1976. Women of value, men of renown: New perspectives in Trobriand exchange. Austin: University of Texas Press. Weiner, Annette B. 1982. Ten years in the life of an island. Bikmaus 3:64-75. Weiner, Annette B. 1983. From words to objects to magic: Hard words and the boundaries of social interaction. Man 18:690-709. Weiner, Annette B. 1988. The Trobrianders of Papua New Guinea. New York: Holt, Rinehart and Winston. Weinreich, Uriel, William Labov, and Marvin I. Herzog. 1968. Empirical foundations for a theory of language change. In Directions for historical linguistics, eds W. P. Lehmann and Yakov Malkiel, 95-195. Austin: University of Texas Press. Whorf, Benjamin Lee. 1958. Science and linguistics. In Readings in social psychology, eds. Eleanor Maccoby, Theodore M. Newcomb, and Eugene L. Hartley, 1-9. New York: Holt, Rinehart and Winston. Whorf, Benjamin Lee. 1940. Science and linguistics. Technology Review 44:229-231, 247, 248. Whorf, Benjamin Lee. 1963. Grammatische Kategorien. Sprache, Denken, Wirklichkeit, ed. Benjamin Lee Whorf, 133-139. Reinbek, Germany: Rowohlt. Wittgenstein, Ludwig. 1971 [1922]. Tractatus logico-philosophicus: Logischphilosophische Abhandlung. Frankfurt am Main, Germany: Suhrkamp. Wittgenstein, Ludwig. 1977 [1958]. Philosophische Untersuchungen. Frankfurt am Main, Germany: Suhrkamp. Wittgenstein, Ludwig. 1980 [1958]. Das Blaue Buch: Eine Philosophische Betrachtung Das Braune Buch. Frankfurt am Main, Germany: Suhrkamp. Worsley, P. M. 1954. Noun-classification in Australian and Bantu: Formal or semantic? Oceania 24:275-288. Wurm, Stephen A. 1981. The possessive class systems in AIWO, Reef Islands, Solomon Islands. Papers in New Guinea Linguistics, No. 21 Pacific Linguistics, Series A, No. 61:181-209. Canberra: Australian National University. Zubin, David, and Klaus Michael Kopcke. 1986. Gender and folk taxonomy: The indexical relation between grammatical and lexical categorization. In Noun classes and categorization, ed. Colette Craig, 139-180. Amsterdam: John Benjamins.
This page intentionally left blank
Index Acquisition of CPs, 25, 32-62, 180-97, 290-94, 296, 298. See also Gender differences in CP acquisition and data on English and German, 195-97 and semantic domains, 291-94 and word classes with CPs, 32, 41, 42, 51, 181 fairy tales and the, 39, 50, 71, 181 in other classifier languages, 193-95 main trends in the, 180, 181, 193, 195, 293-94 order of the, 183-93, 290-94 puberty and the, 62 temporal progress of the, 32, 33, 35, 36, 40-45, 51, 52, 62, 180 Acquisition of nominal classification systems, 13 Acquisition of number distinctions, 13n Acquisition studies in American Sign Language classifiers, 13n, 193 Adams, Karen L., 4, 5, 6, 7, 9, 10, 11, 23, 291, 315, 352 Adorno, Theodor W., 325 Age groups of consultants. See Consultants; Data basis Aims of the book, x-xii, 14, 24, 25, 323-24, 329 Allan, Keith, 3, 4, 5, 9, 10, 11, 13, 14n, 194, 314 Altaic languages, 4 American Sign Language, 4n, 13n, 193 Animate referents and general CPs, 205, 212, 214, 247, 250, 298 Apprehension, 4n Aristotle, 295 Asmah, Jaji Omar, 5, 7, l0n, 14n, 352 Austen, Leo, 318
Austro-Asiatic languages, 4 Austronesian languages, 3. See also Malayo-Polynesian languages Baldwin, Bernard, 14, 26 Bantu, 4n, 194 Barz, Roger, 6, 9, 203, 352n Bechert, Johannes, 313n Becker, Alton L., 5, 9, 10, 11, 13, 14n, 314 Bell-Krannhals, Ingrid, 316, 320 Bellugi, Ursula, 4n Benton, Richard A., 5n, 7, 9, 11n Berlin, Brent, 5, 6, 7n, 9, 10, 11, 12, 13, 18 Berry, J. W., 313n Biersack, Aletta, 319 Bloomfield, Leonhard, 5 Body parts and counting systems, 319 Body parts and reference to kinfolk, 319 Bohme, Karin, 196, 232 Bousfield, John, 215n Boyowa. See Kilivila Broadfield, A., 5n, 313n Bruner, Jerome S., 313n, 314 Budibud, 3 Buhler, Karl, 23, 314 Burling, Robbins, 4n, 5n, 7, 9, 11, 13, 194 Burmese, 10, 11, 194 Bussmann, Hadumod, xii Capell, Arthur, 3, 14, 24, 172, 178, 323 Carpenter, Kathie, 6, 7, 12, 13, 193, 194, 195, 205, 352n Carroll, John B., l0n, 313n Casagrande, Joseph B., l0n, 313n Cassirer, Ernst, vii 371
372 Categorization, 9. See also Classification functions of, 314 philosophical tradition of, 312-14 Central Thai, 12 Chenalho. See Tzotzil Chin, Ng Bee, 4n, 13, 14n, 194, 195, 214, 215n, 291 Chinese, 5n, 194 Cholodovic, A. A., 6 Clark, Eve, 4n, 13 Classification, ix, 5. See also Systems different languages and cultural determination of, 9, 10, 321-22 double, 19, 20, 51, 72, 84 inherent, 6, 8, 11 of classifiers, 8 of the CP system, 198-201 paradigmatic and taxonomic, 9 semantics and, 4, 9 temporal, 6, 8 Classifier languages, 3-5 characteristic feature of, 4 consultants in studies on, 12 data presented in studies on, 12 ethnosemantics in studies on, 10 inventory of classifiers in, 9, 11 language families with, 4 methods and aims of studies on, 12 no color classifier in, l0n the noun in, 4, 5, 6 types of, 4 Classifiers, 4-8. See also CPs action, 8, 9n actual use of, 10, 11, 13 American Sign Language and, 4n, 193 auto-, 7 cannibalizing, 8, 9n, 315 classification of, 7-8, 9n, 198-201 constituent order of the noun phrase and, 4, 5 context influence on choice of, 13 echo, 7, 8 functions of, 6, 11 general, 7, 198, 239 genitive, 9n graphemic, 4 identical, 7, 8 imitative, 8, 9n inherent state, 8 meaning and, 11 mensural, 6, 8 metric, 8, 9n nominal, 8 noun, 8
Index optional use of, 4n origin of, 7n, 352-53 possessive, 9n proper, 8 reduplicative, 8, 9n relational, 9n semantically neutral, 7, 8 semelfactive, 8, 9n sortal, 6, 8 temporal state, 8 true, 8 verbs and, 5n zero-morpheme, 4n Classifier systems, x, 4, 5. See also Systems complex description of CPs in, 25, 327-29 culture-specific definition of, 10, 315-22 development of, 7n, 352-53 features relevant for the structure of, 9, 10 interdisciplinary approach to, 7n semantics and choice of CPs in, 11, 244-83 structure and function of, 9-11 syntax and semantics in, 5 universal status of categories in, 10, 315-22 Cole, Michael, 313n Collins, Allan M., 325 Concepts and CP type choice. See Variable rules, for CP type production Conklin, Harold C, 4n, 5, 7, 9, 23 Consultants, 26, 330-41. See also Data basis; Methodology approaching, 29-31, 296-98 Count terms, 4n CPs (classifactory particles [in Kilivila]), x, 7. See also CP types and adjectives, 17 and demonstratives, 16, 17 and interrogative pronouns/adverbs, 17 and numbers, 17 in Kilivila word formation, 16-18, 22, 31 system of, 8. See also Systems use of the term, 7, 16n CP types, 172-79. See also Language change in the CP system and their frames of reference, 202-27 and their occurrence in recorded texts, 179-80. See also Discourse
Index connecting semantic domains, 283-88 consultants and number of elicited, 131 definition of, 171-72 expected but not produced, 134 produced but not expected, 134,135, 137, 138, 139, 142, 149, 150, 151, 152, 155, 156, 158, 159, 160, 162, 163 produced versus expected, 36-37, 46-47, 56-58, 67-69, 78-80 provisional definition of, 34, 131 CP types/tokens in elicitation test, 32 number of, 32, 39, 40, 41, 52, 55, 56, 63-65 produced by all consultants, 87-130, 132-34, 184-86 produced by consultants in age groups I-V, 35, 36, 39, 40, 41, 43, 44, 45, 50, 51, 52-54, 61, 63, 66, 72, 77, 81, 83, 85, 86 produced versus expected, 134-61 range of, 32, 41, 52, 62, 63, 73, 74, 82, 131-34 word classes and, 32, 33, 39, 40, 41, 51, 52, 60, 63, 72, 73, 85, 86, 168-70 CP types/tokens in overall corpus, 39, 40, 50, 61, 70-72, 80-85, 161-63 nouns and, 166-68 word classes and, 168-69 Craig, Colette, 4n, 5n, 9n, 10, 13, 313 Crystallization, 327-28 Dasen, Pierre R., 313n, 319 Data basis, 31-170. See also Consultants and CP types/tokens produced versus expected, 135-61 characteristic features of the, 31 for or age group I, 32-41 for age group II, 41-51 for age group III, 52-62 for age group IV, 62-73 for age group V, 73-86 for all consultants, 87-134 for the restudy, 301-07, 309, 310, 342-51 in the overall Kilivila corpus, 161-70, 333-38 presentation of, 31-32 Data computation, 31 Data elicitation, 27-31, 296-98. See also Methodology Questionnaire in age groups I-V, 32-86 situation of, 31, 273 word classes and, 31
373
Delancey, Scott, 7n Delimitative markers, 5n. See also Classifiers Demuth, Katherine, 7n, 13n, 14n, 194 Denny, J. Peter, 4n, 5n, 6, 7n, 9, 313n Diller, A. V. N., 6, 9, 203, 352n Discourse, 13, 18, 20-22, 23 Dittmar, Norbert, 326 Dixon, Robert M. W., 4n, 7n, 14n, 215n Downing, Pamela, 5, 11n, 13 Dravidian languages, 4 Durkheim, Emile, 313n Eibl-Eibesfeldt, Irenaus, 18, 320 Ellen, Roy F., 9, 12, 313n Ellipsis, 20-23. See also Classifiers, functions of English, 13, 195, 196 Erbaugh, Mary S., 11n, 13, 14n, 39, 193, 194, 195 Ethnosemantics, 10 Ezard, Bryan, 16n Fairy tales. See Acquisition of CPs, fairy tales and the; Production of CPs, text category and the Fellows, S. B., 14 Fischer, Gero, 7, 9n, l0n Foley, William A., 4 Foucault, Michel, 9, 313 Fox, James J., 325 Frake, Charles O., 9, 313n Freilich, Merris, 12 Friedrich, Paul, 5, 6, 9, 10, 14n, 319 Frontiers of linguistic descriptions, 327-29 Fulguratio, 328-29 Functions of CPs, 16-23 age specifying, 19 discourse, 20-22, 23 ellipsis and, 20-22, 23 gender specifying, 19 grammatical, 16-20 metaphors and the, 19 nominalization, 19, 20, 22 numeralization, 19, 20, 23 verblike expressive, 19, 20, 23 Gandour, Jack, 13, 193, 194, 195 Gegeo, David W., 314 Gender differences in CP acquisition, 33-35, 41, 42-44, 51, 52, 55, 62, 182-83, 187-90, 193 reasons for, 34, 42, 55, 182-83
374
Index
Gender differences in CP production, 33-35, 41, 42, 44, 51, 52, 53, 54, 60-61, 63-65, 72, 73, 74, 75, 76, 85, 131, 166, 182-83, 187-90, 212, 231, 232, 235 reasons for, 55, 76, 182-83, 217, 224, 235, 255, 274 German, 195, 196 Gipper, Helmut, 313 Givon, Talmy, 10 Goffman, Erving, 314, 321 Gomez-Imbert, Elsa, 4n Goodenough, Ward H., 203 Goral, Donald A., 5, 6, 7, 10, 11, 13 Greenberg, Joseph, 4, 5, 6, 14n Guttmann pattern, 31n Haas, Mary H., 5, 9, 13, 194 Hallpike, Christopher, 313n, 317 Harrison, Godfrey, 13, 194, 195 Harweg, Roland, 4n, 9n, 131n Haugen, Einar, 313n Heeschen, Volker, 313 Heraklit, 317 Herder, Johann Gottfried, 313 Hierarchy index, 288, 289. See also Semantic domains, hierarchical order of Hiranburana, SamAng, 9n, 10 Hla Pe, 4n, 5, 6, 11 Hoa, Ngyen Dinh, 4n, 5n, 9n, 10 Hokkien, 195, 214 Hopper, Paul J., 23 Humboldt, Wilhelm von, 313, 327 Hundius, Harald, 4, 5, 9 Hutchins, Edwin, 212, 317 Indo-Aryan languages, 4 In-domain substitution, 243, 244. See also Methodology Intradomain substitution, 243. See also Methodology Inventory of CPs, 16, 24, 26, 172-97, 299. See also CP types no loanwords in the, 16 Jackendoff, Ray S., 4n Jacob, Judith M., 4n, 14n Japanese, 194 John, 313 Johnson-Laird, Philip N., 325 Jones, Robert B., 7n Kaden, Klaus, 10 Kaile'una Island, 26, 31
Kantor, Rebecca, 4n, 13n Kasaipwalova, John, 214 Katz, Elisabeth, 4n Kegl, Judy, 4n Kendon, Adam, 319 Khasi, 315 Kilivila, x linguistic classification of, 3 the numeral classifier language, 4 studies on classifiers in, 14-15 Kiriwina. See Kilivila Klaas, Gerd, 31n Klassifikation. See Classification Klassifikator, 8. See also Classifiers genereller, 8 isolierter, 8, 9n Klaus, Georg, xii Klein, Wolfgang, 31n, 62, 180, 326 Klima, Edward S., 4n Koch, Gerd, 319 Koch, Walter, 325 Koestler, Arthur, ix, 312, 314, 327 Kolner Universalienprojekt, 4n Kolver, Ulrike, 4, 5, 7, 9, 12, 14n Kopcke, Klaus Michael, 4n Labov, William, 11, 203 Lakoff, George, 313, 314, 325, 327 Language change in the CP system, 202-36, 308 multiple classification and, 232-34 opponents of, 228-29, 231 possible consequences of, 236 schoolchildren and, 234, 235 supporters of, 229-30, 232-35 Language, culture, and cognition, 25, 312-22 Language evaluation test, 299-300. See also Methodology Lawton, Ralph, 14, 15, 19, 172, 176-78,. 236n, 308, 323n Leach, Edmund, 313n Leach, Jerry W., 214, 318 Lean, G. A., 319 Lee, Michael, 4n, l0n, 352n, 353 Lehmann, F. K., 7, 11n Lenneberg, Eric H., 10, 62, 313n Lenz, Jakob Michael Reinhold, vii Leon, Lourdes de, 17, 218, 315, 352 Levelt, Willem J. M., 31n Levinson, Stephen C, 13 Linguistic awareness, 228, 231, 232, 235, 300 Linguistic insecurity in CP production, 38, 49, 51, 59, 69, 73, 80, 85, 232
Index multiple classification and, 39, 49, 51, 60, 70, 73, 81, 188-90, 232-34 Lithgow, David, 14, 24 Lloyd, Barbara B., 313 Lobel, Elisabeth, 4n Lonning, Jam Tore, 4n Lorenz, Konrad, 328 Luke, K. K., 13, 194, 195 Lyons, John, 3, 7, 10 MacLaury, Robert E., l0n Malayo-Polynesian languages, 4 Malinowski, Bronislaw, ix, x, 3, 7, 14, 16n, 20, 24, 26, 183, 200, 212, 242, 290, 314, 316, 317, 318, 319, 320, 321 Malmberg, Bertil, 313 Marginal native, 12 Massim, 16n, 318n Mass terms, 4n Matsumoto, Yo, 13, 194, 195 Mauss, Marcel, 313n Metaphors, 19 Methodology, 14, 25-31, 181, 182, 183, 197, 198, 202, 227, 229, 230, 231, 233, 237, 238, 242-44, 259, 275, 282-83, 288, 290, 295-98, 299-300, 320, 321, 324, 325-27. See also Data elicitation; Hierarchy index; Pretests; Questionnaire Miller, George A., 299, 325 Milner, George B., 14 Miram, Helga-Maria, 6, 9, 10, 12, 14n Mithun, Marianne, 7n Mitontik. See Tzotzil Mixtec, 218 Mon-Khmer languages, 4 Mufwene, Salikoko S., 4n, 5n, l0n, 11, 13n, 196 Muyuw, 3 Mythology and CPs, 214-215, 317 Nahuatl, 4n Networks, 325 and their linking elements, 325-26 as "part" within a "whole," 327-29 dynamics of, 326 linear, 326, 327 linear hierarchical, 326, 327 multidimensional hierarchical, 326-27 Newport, Elissa, 4n Niger-Congo, 13n Notation conventions, 17, 31, 32, 87, 131, 134, 135, 161, 237, 243-44, 288, 300, 308. See also Table numbering
375
Noun class systems, 31n Numerativ(e), 5, 6, 7, 8, 9n. See also Classifiers Ochs, Elinor, 314 Off-domain substitution, 243. See also Methodology Orientation, 317 Overgeneralization, 194, 197, 198, 298 Parsons, Terence, 4n Plank, Frans, 4n Posner, Michael, 10 Possessive pronouns in Kilivila, 16n Powell, Harry A., 317 Pretests, 25, 26, 31. See also Methodology Pribram, Karl H., 283, 325 Process of fulguration. See Fulguratio Production of CPs, 25, 197-236. See also Language change in the CP system and context restrictions, 221, 227, 273, 280, 290 dubious and unacceptable, 38, 41, 45, 48, 49, 58, 59, 61, 65, 69, 76, 80, 85, 135, 164, 165, 166 foreign words and the, 60, 84 in elicitation test. See Data basis loanwords and the, 50, 51, 60, 84 repairs in the, 40, 72, 84, 85, 86 speakers decisions for the, 283 text category and the, 39, 50, 60, 71, 80, 82, 84, 85, 221, 227, 280, 290 Protagoras, 312 Prototype theory, 10 Pullum, Geoffrey K., 312 Quantification, 4n, 5 Quantifier, 6, 8. See also Classifiers Quantifikator, 8. See also Classifiers Questionnaire, 27-29, 342-51. See also Language evaluation test; Methodology Quillian, M. R., 325 Rausch, P. J., 6 Ray, Sidney H., 14 Realization of CPs. See Production of CPs Reference to concepts. See Variable rules, for CP type production Referents, xii, 9 Relative frequencies, 244, 275, 282, 300-307, 308. See also Methodology Repairs. See Production of CPs, repairs in the
376
Index
Repeaters, 6, 1, 8, 135, 137, 138, 139, 149, 150, 152, 156, 158, 159, 160, 352-53. See also Classifiers finite set of, 7 Replicability of results, 324 Restudy, 31, 295-311. See also Methodology Role, 203n, 292 Romaine, Suzanne, 244 Romney, A. Kimball, 5, l0n, 11, 12 Rosch, Eleanor, 10, 11, 13, 313, 314 Ross, Malcolm D., 3, 353 Royen, Gerlach, vii, 6, 7n, 10, 16n, 23, 312, 314 Rude, Noel, 4n Sample. See Consultants; Data basis Sanches, Mary, 13, 194 Sapir, Edward, 313 Sapir-Whorf hypothesis, 313 Saul, Janice E., 9 Schafer, Edward H. Jr., 5n, 7, 11 Schiefenhovel, Wulf, 320 Schieffelin, Bambi B., 314 Schleiermacher, Friedrich D. E., 313 Schley, Sara, 4n Schmidt, Arno, 62 Schone, Hermann, 317 Scoditti, Giancarlo M. G., 84 Scribner, Sylvia, 313n Seiler, Hansjakob, 4n, 6, 9n Seiler, Walter, 352 Semantic domains, 9, 12, 25. See also Systems as scripts, pathways, etc., 283, 325 boundaries between, 10 constituted by CPs, 236-42 consultants modification of, 298-99 dynamics of, 14n, 25, 238, 242-90 evaluation of CP types in, 245, 246, 247, 249, 254, 258, 259, 260, 263, 265, 267, 269, 271, 272, 274, 275, 276, 277, 279, 280, 281, 282 features for the definition of, 10 hierarchical order of, 237, 288-90, 296-99, 315-21. See also Hierarchy index universality of, 321 Semantic network. See Networks Semantic scope, 198 Semantics of CPs, 25, 236-90, 315-22 Senft, Barbara, 18, 26, 34, 182, 193, 235, 242 Senft, Gunter, 3, 4n, 6, 9, 14n, 16, 17, 18, 20, 22, 26, 50, 80, 82, 140, 171,
178, 179n, 182, 183, 193, 198, 234, 235, 242, 244, 290, 314, 315, 317, 319, 320, 321, 323n, 325, 326, 352 Serzisko, Fritz, 4, 5, 6, 7 Sesotho, 13n, 194 Shape, 10. See also Semantic domains Silverstein, Michael, 4n Sinclair, Anne, 231 Sino-Tibetan languages, 4 Slobin, Dan I., 322n Smith, Kenneth D., 7 Sociolinguistic variables, 11, 202, 203, 205, 206, 207, 209, 210, 212, 213, 215, 216, 217, 219, 221, 223, 224, 225, 226, 228, 247, 250, 255, 256, 259, 260, 261, 265, 266, 270, 271, 274, 283, 292, 293, 300, 307, 308, 309, 310, 311. See also Status; Trobriand clan system Spitulnik, Debra, 4n Sprachspiel, 326 Status, 202, 203, 205, 206, 207, 208, 209, 210, 215, 216, 217, 219, 221, 223, 225, 228, 229, 230, 231, 232, 234, 235, 247, 250, 255, 256, 259, 260, 261, 262, 263, 265, 266, 268, 270, 271, 274, 283, 292, 293, 307, 311. See also Sociolinguistic variables; Trobriand clan system Stern, Clara, 196 Stern, William, 196 Stolz, Thomas, 4n, 6 Strawson, P. F., 13 Supalla, Ted, 4n Systems, xii, 9 Szent-Gyorgyi, Albert, 312 Table numbering, 24. See also Notation conventions Tatuyo, 4n Tauwema village, 25, 26, 31 Taxonomies. See Systems Taylor, Ronald, 328n Thai, 194 Theory and empirical linguistics, x-xi, 325 Trobriand activities of measuring, 320 Trobriand calendar, 318 Trobriand canoes, 316n, 318 Trobriand clan system, 21, 235, 316. See also Sociolinguistic variables; Status ranking of clans in the, 202, 229 Trobrianders subsistence, 316, importance of yams for the, 320-21
Index Trobriand Islands, x, 3, 24, 26 Trobriand material culture, 316, 319 Trobriand matrilineal society, 235, 317n Trobriand names, 317-18 Tversky, Barbara, 13, 196, 294 Tyler, Stephen A., 9, 197, 294, 313n Tzeltal, 9n, 12 Tzotzil, 315 Unterbeck, Barbara, 5, 6, 12, 298, 314, 325, 352 Vachek, Joseph, 14 Validity of analyses, 324. See also Restudy Variable rules, 244 correct prediction of CP type production by, 309-11 for CP type production, 244-82 Variety grammar, 326 Vater, Heinz, 226
377
Veer, Kees van der, 31n Verbal derivational classificatory affixes, 16n Verhaar, John, 4n Vietnamese, 9n Vollmer, Gerhard, 313, 314, 317, 326 Wallace, Ron, 325 Wassmann, Jurg, 319 Watson-Gegeo, Karen Ann, 314 Weiner, Annette B., 182, 183, 232, 234, 242, 290, 317n, 319, 320, 321 Whorf, Benjamin Lee, 313 Wittenburg, Peter, 31n Wittgenstein, Ludwig, 236n, 323n, 326 Wundt, Wilhelm, 314 Wurm, Stephen A., 16n Yucatecan Maya, 10, 12 Zahleinheitsworter. See Classifiers